Posted on August 7, 2023 by rtushnet

Hall v. Marriott Int’l, Inc., No. 19cv1715-JO-AHG, 2023 WL
4417265, — F.R.D. – (S.D. Cal. Mar. 30, 2023)

Previous
opinion in this case about allegedly deceptive/inadequately disclosed mandatory
“resort fees.” The court ends up certifying an issues California class, but
rejecting the damages methodology for calculating class damages.

Marriott’s booking process shows prices starting “from” a
particular rate:

 

“from 420 USD/night” ad

Clicking on “view rates” gets further information:

 

still says 420 at bottom but now there’s a message at top in blue

If a resort fee applied, a notice would appear at the top of
the screen in blue, bolded text that is outlined by a blue box, stating “Please
note,” “USD 35 daily destination amenity fee will be added to the room rate.”
The second page in the booking process also reflects the available rooms (e.g.,
queen or king room) and corresponding rates, without the resort fee included.

Once a consumer selects a particular room, they arrive at
the third page in the booking flow:

 

at last the fees are at the bottom in the total

The subtotal shows a charge for “USD/night” and a separate
charge for “USD taxes and fees,” which includes the resort fee. A more detailed
breakdown of the charges showing the amount of the resort fee is available if
the consumer clicks the “Summary of Charges” dropdown box:

 

at checkout the total is correct

Plaintiffs alleged both that Marriott inadequately disclosed
resort fees on its own website and on the sites of third party online travel
agencies. They brought the usual
California claims.

The court dismissed all equitable claims because plaintiffs
lacked standing to seek injunctive relief, and the court lacked jurisdiction
over the equitable claims because damages hadn’t been shown to be inadequate.

CLRA: Plaintiffs had two theories of deception: (1) bait and
switch based on the initial ad for the room rate that was unattainable without
paying the resort fee, and (2) inadequate disclosure of the resort fee—the blue
box was in smaller print, a different color font, and at the top instead of
next to the price information at the bottom. The “taxes and fees” disclosure
was allegedly confusing because it does not explicitly specify “resort fees”
and causes consumers to believe that the fees are entirely government related.

Theory 1: On the one hand, “explicit and conspicuous
qualifying disclosures can render allegedly deceptive statements non-deceptive
as a matter of law,” but, “[e]ven when there is no question that the initial
deception was cured by a later disclosure, a company can still be liable for
deceptive practices that cause consumers to be lured in and ‘swept up’ in the
buying process.” The latter situation occurs when deception causes “consumers
to invest significant amounts of time and become so swept up in the buying
process that later qualifying disclosures cannot cure the original deception,”
and is exemplified by a brick and mortar store advertising 40% off in the
window, which led consumers to enter, shop, decide to buy, and stand in
line.  By the time qualifying disclosures
were made, the consumers were “invested in the decision to buy and swept up in
the momentum of events.” The plaintiffs had spent 40 minutes or more shopping,
waited in lines of 15 or more people, and felt pressured to purchase once they
reached the front of the line due to embarrassment.

By contrast, 
“Marriott’s disclosures regarding the total price of the hotel stays are
conspicuously disclosed by the end of the internet booking process.” The
initial “from” (which I think is deceptive, since it’s not attainable) was
followed by disclosure of additional fees multiple times. The plaintiffs didn’t
invest significant time in the purchase process—at most 10 minutes total, and
the higher price appeared on the second page of the booking flow. The
plaintiffs were aware that the price increased throughout the booking process.

However, there were disputed issues of fact regarding the
deceptiveness of Marriott’s booking process and the adequacy of its disclosures.
“Unlike its disclosure of the total price, Marriott does not repeatedly and
conspicuously disclose resort fees throughout the booking process.” There were
genuine and material factual disputes about whether consumers would notice and
understand the disclosures due to their font size, color, and placement within
the context of the entire transaction. “Indeed, Plaintiffs point to survey
evidence, that approximately 50% of consumers do not notice these disclosures
and do not know they paid resort fees.”

Negligent misrepresentation failed because nondisclosure is
not the same thing as making a false statement.

Marriott could not be held liable for allegedly deceptive
statements on third-party sites. There was no evidence that Marriott controlled
how they presented resort fee information. Under the CLRA, vicarious liability requires
evidence of “personal participation in the unlawful practices and unbridled
control” over those deceptive practices.

Marriott’s class action waiver defense failed as to the
named plaintiffs, but could still be raised as to unnamed class members.
Marriott bears the burden of proof on this issue, and pointed to no evidence in
the record that the named plaintiffs had notice of Marriott’s terms and
conditions and assented to them; even the terms and conditions they purportedly
agreed to were absent from the record. However, Marriott sufficiently preserved
and maintained its right to assert a class action waiver defense against the
proposed class members.

Marriott also failed to show a lack of reliance. A plaintiff
can show reliance by showing that the misrepresentation or omission at issue “played
a substantial part … in influencing his decision,” or by showing that the
misrepresentation or omission was material. A plaintiff need not show that
“[the challenged] misrepresentations were the sole or even the decisive cause
of the injury-producing conduct.” Reliance for omissions is satisfied if the
plaintiff demonstrates that “had the omitted information been disclosed [the
plaintiff] would have been aware of it and behaved differently.” There was a
triable issue, despite testimony from the lead plaintiffs that they chose
Marriott on factors unrelated to resort fees, such as total price. Although
this testimony may demonstrate that resort fees were not “the sole or even the
decisive cause” of plaintiffs’ decisions to book their hotel rooms, it does not
establish that resort fees were not a “substantial part” in their decision
making.

The court found that it couldn’t certify a nationwide class,
but Rule 23(a) was satisfied with respect to a California class. The Rule 23(b)
problem was predominance. First, the proposed class definition encompassed a
“potentially significant number” of unharmed consumers who saw the disclosure,
which might be about 50% on plaintiffs’ own survey. Second, the proposed
damages model likewise didn’t distinguish between injured and uninjured class
members. The expert calculated the total amount of resort fees paid during the
class period, but not all of that would reflect injury.

Still, the court certified a liability-only class. “Most
critically, all of the essential liability elements of Plaintiffs’ CLRA and
common law fraud claims can be resolved on a classwide basis using Plaintiffs’
common evidence.” This would also advance judicial economy, given the small size
of individual recovery and large numbers of individual class members.

from Blogger http://tushnet.blogspot.com/2023/08/hall-v.html

Posted in Uncategorized | Tagged , , | Leave a comment

IPSC Closing Plenary Session

Posted on August 4, 2023 by rtushnet

An Author/Reader Conversation about Jessica Silbey, Against
Progress: Intellectual Property and Fundamental Values in the Internet Age (2022)

Robert Brauneis: Three layers—(1) object of discovery:
creator and innovator accounts, concerns, what kind of conditions support/hinder
them in their work; (2) thinking about IP rules, particularly © for
photographers; (3) economic & social vision; equality, dignity, privacy,
etc are contrasted to hierarchy, subordination, exclusivity, precarity,
commercialization. Changes to IP rules on their own in our world may have little
effect on whether we go towards the good vision or the bad—cloud computing,
network effects that cause convergence on a small number of intermediaries, and
important tech that is protected by secrecy/real and personal property
law/employment agreements. IP loosening could not have much effect/even help
strengthen corporate consolidation. Norms/inculcation of values and other
fields of law like antitrust might be more effective.

Questions about whether creators would agree—Lynn Goldsmith objects
to Warhol’s use; maybe that’s an effect of being an unusual use compared to how
photographers usually see their works adapted in art. Maybe it’s a sense of
breach of contract, but the book describes more tolerance for reuse.

Silbey: ©, TM or patent becomes a device to assert certain
things about yourself or your work. Insofar as people like Lynn Goldsmith or
the ACLU are using IP to argue for certain justice goals, what IP is and what
it’s for may be changing, even if the text doesn’t change.

Deepa Varadarajan: IP is a vital terrain for contesting fundamental
values. As fiction author, struck by community norms among creative
communities: creators routinely avoid the constraints of IP regulation and err
on the side of more promiscuous sharing—fairer uses. But also less willing to
tolerate uses “not in the same spirit.” Should IP law try to incorporate more
norms, especially when defining a community is getting harder and harder? (Compare
BookTok—are readers part of the fiction community?)

Trade secrecy: mixed up with commercial morality and relational
duties, not just incentives—is this more receptive to incorporating the broader
set of values described the book? Also intersects a great deal w/contracts. Contracts
imposed on people w/lower bargaining power can get rid of the limits on trade
secret doctrine; this is also a theme of the book—form contracts can undermine
the rights and recognitions that creators seek. Pro photographers agree to
onerous contracts from longstanding clients in order to retain them. NYT
theoretically lost Tasini, but led NYT and others to require fee-free transfers.
Giving more rights to authors didn’t give them more bargaining power against
aggregators.

Michael Burstein: seems like disorganization is an issue—Conde
Nast being able to impose terms on scattered photographers seems like a reason
for discontent w/ private ordering

Rebecca Curtin: Important difference from tech upheavals of
past. In manuscript to print, you can find readers anticipating what print will
facilitate—the concept of an authoritative edition, the professional editor;
readers wanted these things before the tech offered them. In Silbey’s book, the
tech has lapped creative communities and begun to unravel norms central to
creation and dissemination rather than coalescing them.

One lesson: Threats to privacy are threats to communities
and practices that sustain creativity. The “clean air and water” of culture are
at risk.

Michael Burstein: Presence in narratives of transactions as
central—IP might be in the back seat. How the values come into play is less in
creation and definition of IP rights and more in their flow, transfer through
ecosystems. Equality: in discussion of equality, in Tasini, Roche, and even
Kirtsaeng—the first two seem to have resurrected the romantic author not as
creator but as transactor: each case claims to protect small creator/inventor,
with little awareness of practical effects. IP law of creation isn’t telling
the whole story.

Open arrangements/commons-based accounts are largely devoid
of law and more about institutional structures/governance and self-ordering
that is more than just in the shadow of the law.

Institutional precarity: Declining trust in markets, not in
IP law. Creators experience these practices as coercive, hostage-taking, and
the effect is on attitudes towards market structure. That’s the domain of other
kinds of laws, like antitrust. Those markets obviously depend on definition of
goods sold in them, and that can’t be ignored.

Public interest is often missing in anti-discrimination
context like Eldred and Golan, and anti-subordination analysis like Tasini and
Roche. Individual interest v. public interest opposed in those cases and in
Kirtsaeng. Wonders if the baseline in IP renders issues of civil equality more
difficult. It’s easier to identify superior moral claim in race/gender
discrimination than in the public domain. These are policy choices, pushing IP
closer to regulation than to property regimes.

Rosenblatt: power imbalance and the importance of collective
action have become more central to our lives/scholarship. Ability/inability of
creators to work collectively seems a recurring theme in addressing power
imbalance. What does that mean for us as IP thinkers?

Josh Sarnoff: end of liberalism v. paternalism—we’re seeing
that play out in IP. Paternalism in IP can be things like preemption overriding
contractual waivers/overrides.

Silbey: wanted to reaffirm the idea of the public, not the
public domain—we are all in this together, interdependent.

from Blogger http://tushnet.blogspot.com/2023/08/ipsc-closing-plenary-session.html

Posted in Uncategorized | Tagged , , , , | Leave a comment

IPSC Breakout Session #5 Platforms & Interfaces/IP Enforcement

Posted on August 4, 2023 by rtushnet

Xuan-Thao Nguyen, Tech Bros, Social Media, and the End of IP
Financing?

Ideas are nothing without financing. Banks will not give
loans secured by patents; banking law constraints. How was Silicon Valley Bank
able to make loans to startups before its demise? Piggyback on VC due diligence,
valuation. Funding round provides money to pay back loan; SVB also took a
warrant on the startup itself—the right to purchase shares. If the value of the
startup goes up, the value of the warrant comes up.

What will happen now that SVB collapsed? Operated in US,
Israel, China. Note that access to funding was never equitably distributed for
women and minorities. Nonbank lenders may come in, but problems are associated
with that. Nonbank lenders demand higher fees and interest rates b/c their
money comes from investors, not depositors. Bank loans are typically much
cheaper. IP is considered very risky for loans—not good collateral.

Felix Wu: What should happen?

A: senior consultant for World Bank: banking law is not
going to change. Banking focuses on deposits; we have to protect the
depositors. The loans rely on quantifiable collateral. Lessons from China:
patent office can work with bank and experts to value specific IP, issue a
certificate, and the bank can then rely on that. Insurance policy. Can also
have government provided guarantee.

RT: TM as collateral?

A: Has paper in Houston LR: an established company w/royalty
stream can rely on them as accounts receivable. Even a new company, if they
have customers ordering, the bank can use that as financing based on accounts
receivable. The byproducts of the TM can be financed. But IP, including TM, are
not viewed by banking sector as the type of corporate assets on which they can
rely for lending.

David Stein, Rethinking IP Incentives Following a Process
Shock: Lessons from Online Consumer Services

Software IP broke about 10 years ago and no one noticed. First
divide in protecting implementation but not interface, for
competition-based/network effect-based reasons. The problem is that this system
fell apart around 2008-2012 when smartphones came out. Once I have a phone
connected to the internet, I want my software available to me everywhere.
Everything moved into the cloud. No need to bear all implementation costs up
front, as there was before; can add or remove features over time, handle edge
cases that develop. At the same time, the risk of copying essentially went
away. Access to market is now about having enough computers to run the software
for all your users. So now, we have lots of protection for implementation where
we don’t need it and no protection for interfaces where new entrants can easily
be copied.

It’s worse than that b/c of the difference b/t disruptive
and incremental innovation. Totally new product: risks cannibalizing existing
market; risks entire portfolio (example: Google Buzz—still in a consent decree
w/FTC). That doesn’t exist for startups. That means small companies largely introduce
disruptive innovations, while A/B testing of small changes favors
size/economies of scale. Thus if a large company copies a small company’s
innovation, it can accelerate faster than the small company and make it much
better.

Recommendations: dump most precedent. Some need in installable
software world, but if we’re worried about concentration and innovation in
online spaces, the scheme we have doesn’t quite work. Like to see more
protection for disruptive interfaces. Problems exist w/lasting IP rights in
interfaces, b/c right to stop people from distributing software does create
barriers, so misappropriation and unfair trade practices may be more suited.

Fred Yen: These interface problems have existed since Lotus
v. Borland; pick a few examples that people can get their hands around,
especially for a general audience.

A: you have to be a large incumbent before there’s a
community depending on Lotus. Big difference is the direction of copying. If Lotus
controls the market and people are dependent on it, then others need to support
Lotus’s interface to compete. But that’s not the direction of copying he’s
seeing—cool new interfaces by startups are copied by Meta and Google and
Microsoft. That’s more problematic.

RT: Why is IP the solution space? Why won’t the big
companies just acquire and kill startups unless there’s a structural remedy?
Meta knows how to write a big check.

A: Cares about activity levels; ruinous copying decreases
the incentive to try to get bought out. Thinks there’s too much copying—why would
anyone buy when they could copy? [And yet they do buy, a lot.] Structural remedies
mean that there’s no hope of being acquired which depresses innovation.

Jessica Silbey: Give a menu of options: reverse confusion
protection for unregistered trade dress; improvement patents; etc. Much better
for doctrinal gravitas.

A: could have FTC exercise its §5 authority; startups won’t
use the law b/c they can’t survive the length of time it takes to bring a case.

Wu: not convinced there isn’t © protection for many
interfaces; the lock-in effect may override that in a market dominance
situation but that’s different from there being no protection.

Ben Depoorter, Copyright Small Claims Litigation: An
Empirical Analysis

Goal: stick to infringement. Time frame ongoing; 428 cases
so far. Reproduction and public display are the main alleged rights violated
(with lots of overlap); only 71 derivative work claims. Smaller cases, under
$5000, are more streamlined—only 1 Board member evaluates; 163 of the claims
choose this. Statutory damages requested 22% of time. 71% pro se, 28% represented.

Few repeat players, but Joe Hand live entertainment is—business
model is working with bars etc for live entertainment broadcasts. 7% optouts, but
note that you don’t need to opt out if the claim is noncompliant.

Little evidence of trolling, opt-outs; many noncompliant claims.

Next steps: identify trends, decision analysis, damage
awards, types of noncompliance.

Silbey: also look at who asks for the smaller awards.

Q: maybe optouts are lower because of risks of fee-shifting
if the defendant makes a big federal case out of it. [registration status makes
a big difference here]

A: generally, there is a selection effect—wants to track
that over time. There are a lot of considerations; the opt out notice is an ad
for the board—the CO warns about what can happen in federal court.

RT: asked about registration status—which would interact
w/risk of fee shift in federal court. I just checked and the CCB does not explain the effect of a
timely registration in federal court in its list of considerations about
whether to opt out or not.

Number one consideration is what makes initially
noncompliant submissions not try to correct.

Kristelia García, Imperfect Enforcement

What makes people underenforce—retweet fan videos instead of
shutting them down? Mechanisms and benefits are similar in IP and outside.
Unlike public enforcement discretion, private enforcement discretion/selective
enforcement is not well studied or theorized.

Taxonomy: private enforcement can be selective, delayed,
algorithmic; private forbearance can be nonenforcement by default or
intentional; intentional can be ex ante or ex post. Comedians suing Spotify;
tattoo artists suing videogame companies. Questions of reliance/implied license
can arise from delayed enforcement.

Intentional nonenforcement: foreclosures are an example
where there is nonenforcement, disproportionately in wealthier neighborhoods.
Video game developers who allow piracy in hopes that users will buy in game
content.

As with other forms of private ordering, selective
nonenforcement does raise some concerns. Propensity for bias is obvious as
w/mortgage enforcement. Sometimes it’s economically rational—wealthier people
are better credit risks, so maybe they can get their accounts current and take
advantage of loan modifications. Lack of transparency about reasons.

Advantages: decision in hands of party best positioned to
evaluate. Valuable info to lawmakers about what industry cares about and what
it doesn’t. Not every wrongdoing results in a loss.

Lemley: caution about any suggestion of use it or lose it—leads
in TM to bullying and threats of lawsuits that are unnecessary and harmful.
Larger issue: we have this nonenforcement b/c we’ve created an overbroad set of
legal rights such that infringement goes on all the time. So we recognize that
lots of copyright infringements are not worth enforcing, either not harmful or
even helpful. If we’re not ok with that, think about narrowing the scope of the
enforceable right.

Xiyin Tang: what’s the difference b/t delayed and intentional?
Petrella seems to involve elements of both.

Alex Roberts: how does this play into a monitoring strategy?
Aggressive monitoring v. sit back and wait to see what comes to your attention.

A: yes, and algorithms also come into that. Some pass it off
to algorithms.

from Blogger http://tushnet.blogspot.com/2023/08/ipsc-breakout-session-5-platforms.html

Posted in Uncategorized | Tagged , , , | Leave a comment

IPSC Breakout Session #4 Innovation/Copyright

Posted on August 4, 2023 by rtushnet

Room 204 Christopher Buccafusco (w/ Joseph Blocher),
Firearms, Innovation, and Regulation

How do law and markets affect the pace and direction of
innovation for firearm related safety in the US? Costly inefficiencies in
supply and demand.        

Virtually no one thinks the US has the right amount of gun
violence, and the firearms industry has been and continues to be enormously
innovative—AR-15 customization, ghost guns. But innovations to make the guns
themselves safer have failed to appear. E.g., smart guns: thousands of people
are killed/injured each year by guns fired by someone other than the owner. 1
of 6 police officers shot are shot by an officer’s own gun. 250,000 guns stolen
annually and disproportionately used in crime. Shootings by children, suicides,
etc. 2016 survey says 59% of Americans would be willing to buy a smart gun,
including 56% of political conservatives and 4 of 14 gun owners.

Since the 1990s, manufacturers have been producing
functional prototypes for user authentication of handguns, shotguns, and other
long guns—facial ID/biometrics, codes, etc. But none reached the market despite
millions in R&D from National Institute of Justice. Colt and Smith &
Wesson had functional prototypes by 2000, followed by dozens of startups. But
only last week did Biofire offer the first public sale of a smart gun.

VCs: Liberal, not interested in funding guns; more
interested in software than hardware.

Demand: no buy-in from institutional investors like police
forces; some purchasers are deeply hostile to smart guns b/c they fear gov’t
will come for all other guns. Partly in response to NJ’s 2002 law—once there’s
a smart gun, manufacturers have to switch to it w/in 30 months, though NJ
backed off and just required retailers to stock it, but still infuriated gun
rights advocates who boycotted Colt and Smith & Wesson who then got out of
the market entirely. Established firms pulled out of the market, so startups
can’t expect to be acquired and must go all the way to direct sales; there’s
also fear of tort liability. Biofire isn’t submitting for registration in NJ to
avoid triggering law.

Microstamping: tech that imprints a gun’s serial number on
discharged rounds. 2007: Cal. required firms to included microstamping once DOJ
certified that there was no patent on it; challenged under Bruen and dct
overturned the law b/c it prevented people from buying state of the art
handguns/not consistent w/historical tradition which was light on
microstamping. Cal. isn’t even appealing the loss on microstamping.

Limits on internal innovation created opportunities for
external innovation on safety (which is probably bad)—e.g., installation of
Shot Spotter all over. $21 million in Louisiana to “harden” schools, half a
billion in Texas; schools designed with curving walls to decrease damage done
by active shooters; bulletproof backpacks.

Sometimes external environment-level innovation is more
efficient: ramps v. stair climing wheelchairs. But many external innovations
come with huge social costs.

We don’t have solutions; everything sucks. Maybe: subsidies
for smart gun purchases? Institutional commitments from institutions like
police forces?

Q: Europe?

A: some innovation does come from Europe. Is there demand in
these other countries?

Betsy Rosenblatt: one story seems to be that
innovation-forcing laws can inhibit that. Is that unique to this field?

A: No—seen something similar in disability as well. Tort
tries to get people to make safer products to encourage innovation, but also
leads to anxieties about creating new stuff b/c it risks litigation v. doing
what everyone else is already doing. Boycotts are inhibiting Biofire—they have
had to go fully direct to consumer b/c dealers won’t stock them for fear of
boycotts.

Q: federalism story?

A: this is part of the challenge—innovation folks usually
don’t have to think about public law and state v. federal.

Q: what about the military, which seems to be missing from
this story? Probably a big buyer w/an interest in having still-lethal weapons
that limit friendly fire and suicide. Where are they?

A: we are looking for an answer. They were very interested
very early with Colt and Smith & Wesson. These are professionals who know
how to handle guns, so the safety needs may seem less pressing, though that’s
probably wrong. It’s become so politicized, and military is disproportionately “gun
rights” folks. But senior leadership could decide to prioritize safety (maybe
with negotiating not to trigger the NJ law).

Q: VCs are less liberal than people think, and politics take
a back seat to economic opportunities, so why not?

A: the big problem is exit—Colt and Smith & Wesson don’t
want you; you have to be Tesla and go to production. But the story that many of
these people are telling is “we can’t get funding.” Maybe VCs don’t believe in
the tech; the story: there’s money for mental health, victims, and school
hardening, but people don’t want to be in the gun industry.

RT: [So one factor that might be silent here is
outsourcing/contracting: b/c the military no longer makes its own stuff and
seems institutionally incapable of imagining that it might, it is dependent on
outside contractors, and if they won’t do it, too bad. That seems bad for
reasons beyond guns.]

Mark Schultz, Video on Demand Services: New Frontiers in
Regulation of Cultural Policy, Industrial Policy & Copyright
Streaming took off, and regulation was close behind, motivated by concerns
about culture. New cultural policy in Australia, 2023: fear of voices being
drowned out. Unlike free over the air TV, no requirements to make Australian
content available.

Rising wave of interventionist cultural policies: European
AV Services Directive, Australia, Canada have passed regulation and others are
on the way. Argument: likely to fail both as cultural and industrial policy.
Building on other work on cultural and economic policy by Pager, Park &
Messerlin.

A better way: decentralized policies that promote local
capabilities—the strong preference for local content can be competitive. Korean
success story.

Two models of cultural policy/industrial policies. First,
interventionist: on the content side, subsidies (French film industry), content
requirements, language requirements. Industrial policy: local content quotas;
local investment of profits (streamers must invest percentage of revenue, not
profits, locally); local production requirements for location, personnel, financing;
terms of trade (retention of copyright, exclusivity limits—streamer can only
have license for limited amount of time, investment limits on how much
streamers can invest in local companies); import quotas; screen quotas;
prominence requirements (local content must rise in search results).

Second, decentralized—there is no country that is purely
noninterventionist; everybody does something. But broadly, market based, focused
on private investment, content neutral; largely hands-off except for granting ©.
Some countries like Korea invest in building creative skills and technical
skills, building studios and other infrastructure; tax breaks; promotion,
marketing, and other related capability policies.

EU AV Services Directive requires streamers to include at
least 30% “European” content. Permits member states to require re-investment of
streaming revenue locally. France has required 20-25% reinvestment, Italy
considering similar marks. Regulating terms of trade also permitted, including ©
ownership/exclusivity/restricting investment in local productions.

Canada, Bill C-11 passed. Regulations in progress: local
content, if similar to broadcast will be 35-50%. Local production requirements:
not enough to film it here. Must have Canadian producer making decisions; a
point requirement where you get points for, say, screenwriter, which leads to certain
market distortions.

Hasn’t worked well as cultural policy and thus fails as
industrial policy. Sean Pager’s work: as France increased subsidies, its share
of its own box office relative to American share went down. The argument has
been that the French had incentives to create content based on guaranteed
subsidies so there was no incentive to create material that was appealing to
audiences, especially in comparison to American films. The subsidy trap: the
bureaucrat is your audience, which leads to a certain type of filmmaking (not Scorcese
or Spike Lee; willing to take bureaucrat’s suggestions). Censorship isn’t the
main problem—even when the cultural bureaucracy is insulated from politics, the
office culture has its own office politics and may not be interested in what’s
appealing to the local public. There are also cronyism and quota problems: when
you have a quota, people may take advantage of that to make quickies on the
cheap w/low production values.

The Emily in Paris problem: The dodge where you make the
content in the country but not for the country. The Falcon & the Winter Soldier—set
in Eastern Europe but about who is the right person to have Captain America’s
shield.

Distorting local investment: local filmmaker complains it
makes it harder for locals to compete. Streamers are paying local actors more
than local producers—maybe that’s good but it does divert from local-inspired
content. Can pigeonhole locals/block them from opportunities. Margaret Atwood:
book by a Canadian, filmed in Canada, but it didn’t count b/c scriptwriters
weren’t Canadian.

Korea as success story: Language unique to Korean peninsula;
relatively unique culture. First swept through Southeast Asia, Japan, China;
then US, Latin America, Europe. $12 billion/year in exports, plus soft
power/tourism.

Decided in 1993 to focus on culture as strategic sector. Indirect
support: tax credits and incentives for private investment, including
micro-investment; pushed chaebols to be involved, which they were until the
financial crisis when most spun off those parts. Direct support focused on infrastructure
and human capital: production facilities, training, export promotion. This is
the model with the fewest unintended consequences. It’s lowbrow/mass culture,
sure. Most countries do mild subsidies to preserve certain forms of culture;
but media sector shouldn’t be dependent on those subsidies, and ultimately time
tells what it is highbrow or lowbrow.

To avoid the mistakes of interventionist policies of the
past, national governments should promote cultural industry capabilities, but
avoid picking winners in ways that make creative industries complacent.

RT: I find this convincing but I’m interested in what a
French bureaucrat would say in response and your answers.

A: France would say: We make great stuff. We make real art. True:
The Francophone Africa film industry has produced beautiful movies, but no
Africans ever saw them: Nollywood is more popular and tells “African” stories. French
might also point to the fact that, in countries where they dropped controls,
American movies flooded in—as in Mexico, where film industry struggled. Mexican
gov’t didn’t do Korean-style policies, though it did promote telenovelas.

Revealed preferences: if people don’t go see it, does it
matter?

Q: why not delegate to experts about what would be good, not
necessarily popular?

A: experts tend to have their own strong preferences. German
cultural bureaucracy would finance either old German operas or really
avant-garde productions.

Rosenblatt: What’s the role of unions?

A: good question—maybe some interaction.

Rachel Landy, The Innovation Void in Downstream Content
Markets       

Music: Same product, at same price, from three main
companies in our lives plus Spotify. Live online TV has more variation in
price, channel options, etc. What about a $4.99 monthly for nothing but
catalog, no playlists? What about a jazz service or a metal service? Record
industry’s role in suppressing innovation. Labels’ conditions prevent
innovation. High concentration—3 dominant labels with must-have catalogues.
Each can veto an entire business. They are complementary oligopolists: Cournot
complements—you get even more market power and leverage. Coupled w/desperation
to get back to pre-digital levels of control. You see ratesetting and other key
license provisions.

Each label enters into an agreement “independently” with
each service, but there are standard terms. A large up-front minimum guarantee
payment, often in the hundreds of millions. There’s a revenue share for recoupment
against minimum guaranteed; the revenue shares have most favored nation
principles, which allows them to know how the other labels are pricing. Labels
keep any overage, known as “breakage,” and it’s unclear how much if any is
shared.

Super-narrow © license. Services have to get permission from
each label for any new feature or functionality that invokes the catalog, and
again this facilitates information sharing.

Trust: repeat players, reputational sanctions, reciprocity—deter
innovative options. Result: higher end-user prices. Barriers to entry, and less
innovation by incumbents b/c so much is being extracted: 55% of revenue. Also harms
indie artists; incentive to promote major label content to recoup the
guarantee. The labels have seen their own costs go down—no pressing records;
the services pay for the infrastructure.

Solutions? Consent decree frameworks; statutory licenses;
antitrust law reform against tacit collusion; MFN clause ban. Incentives to
defect? We could tax the breakage that can’t be tied back to any content on the
service. Could tax the surplus made through the MFN; might encourage labels to
drop guarantees to a level where they could actually be recouped. Transparency of
parallel contractual provisions might also help.

Tang: There’s more innovation than you say in music—you’re
only talking about premium streaming services, but not iHeartRadio, free
Pandora, Amazon Prime bundling.

A: for webcasting, there’s a statutory license with some
protections built in though they also limit innovation. Those are also
controlled by the labels and subject to more restrictions than the premium
services are b/c the labels want to funnel people to the paid subscription.
There are other parts of the industry where innovation is flourishing—where the
labels can’t do this—TikTok or YouTube UGC.

Kristelia Garcia: the tax thing is intriguing—are there
similar examples?

A: tax as a tool we often look to for encouraging/discouraging
behavior; cigarette and other sin taxes. Congress loves to amend tax code and
not so much ©.

Blake Reid, Copyright’s Periphery

Copyright on a dying planet. Looking at 1201 triennial
rulemaking: 15 years of trauma before the Copyright Office discussing far
ranging areas of law and policy: environmental regulation, disability rights,
medical devices—far from concerns of ©. We’re trying to do serious policy and
somehow we’re funneling it through the distorted lens of ©. © routinely infects
policy areas outside its core of incentivizing creative works. The 1201 review
illuminates the problem.

1201 creates paracopyright liability for circumventing tech
protection measures that control access to © works. There’s no protection for
circumvention aimed at noninfringing or fair uses in most circuits.

2021, there were 21 distinct exemptions, including
traditional categories (motion pictures, video games), but a lot of literary
works as computer software. Worth emphasizing that a lot of these exemptions
are not new, but have persisted across multiple rounds of rulemaking. They show
the © periphery.

Intended beneficiaries are often small/individual—film critics,
documentarians, disability services providers, people w/disabilities, farmers,
repair techs; often public-facing, they often care about complying with the law
and require degree of legal certainty; they often produce public goods like
privacy, education, agriculture, data security.

Many of these uses are functional and uncontroversially
noninfringing: functional uses and modifications. Unlocking, jailbreaking
involve only glancing uses of protected works and are incidental to the use:
the fact that you need to use the software on a tractor engine to repair the engine
is just incidental. Facilitates uncontroversial uses like reading that might be
required by other laws like ADA. Or exposes/tests vulnerabilities of TPMs and
software—used to evaluate and diagnose software. Many uses are noninfringing
but not subject of direct case law.

1201 requires Copyright Office, which doesn’t look at fair
use particularly charitably, to determine that uses are likely noninfringing.
B/c these uses and users are public facing and often chilled ex ante, there is
often no case law on point.

Why do rightsholders object to exemptions? The review is
really contentious despite the noninfringing nature. Objection 1: speculative
abuse of exemptions—encourage infringing behavior adjacent to but beyond the
bounds of the exemption. Across decades, no instance of this has been
identified where a bad actor purports to rely on an exemption. Objection 2:
non-© policy motivations for regulating: 1201 is a proxy for other policy
issues like DOJ treating 1201 as belt and suspenders for CFAA defense against
hackers, as if Russian gov’t cares. Concerns about vehicle modification violating
pollution regulations; FDA worrying about medical devices. Explicitly beyond scope
of © and institutional context where Copyright Office isn’t capable of evaluating;
it tries to do so under 1201’s catchall provision. Why is the Librarian of
Congress in charge of deciding which cellphone you can use? It’s supposed to
consult w/NTIA, but routinely rejects NTIA’s recommendation.

Objection 2.5: non-copyright policy interests of TPM
deployers/rightsholders: disclosing security flaws might be embarrassing; right
to repair might allow independents to compete with authorized repair.

Objection 3: non-copyright micromanagement of user/circumventor
activity. That’s a result of a sense of entitlement from exemption opponents to
control how users behave. CO builds a miniature regulatory scheme into the
exemption, e.g. for text and data mining—security practices and accreditation
requirements for researchers.

What could we do to wall off the periphery?

Lower the bar for securing/renewing/expanding exemptions; encode
more in statute.

Cover development of tools.

Eliminate 1201 or require an infringement nexus.

What would bear on copyright more broadly?

We should consider specific exemptions/limitations for categories
of users likely to engage in the production of public goods, knowing their uses
are likely to be especially sensitive to liability risks; specific exemptions
for functional uses; new institutional contexts for assessing fair use ex ante—declaratory
judgment attempt in CASE Act was an unsuccessful but interesting trie.

More muscular policymaking in non-copyright congressional
committees—don’t defer. Even where there are complex fair use and doctrinal
issues, AI is an example where the policy equities are far beyond © and fair
use’s capacity to address, like labor and privacy.

Zahr Said: Why are you conceding that this is on the
periphery? Post-colonial theory makes it feel like a concession.

A: experience of triennial review, which represents a long
dedication from a lot of communities who band together. All that engagement
with the core gets indifference from actors who are at ©’s core. They’ve tried
but there is so much skepticism and distrust for exemption proponents.

Charles Duan: it’s cheap to use DRM and get the value of
excluding people from an entire device. “Cheap Exclusion”—relevant paper. Value
disconnect.

A: I’m skeptical that they’re really getting much value out
of this [Duan and Rosenblatt: They think they do!] Some of their claims are
just ridiculous—using 1201 to stop password sharing on Netflix is never going
to happen. There are no 1201 lawsuits about that, or about anything really.
Used for B2B disputes. But it’s cheap to send someone to the CO.

RT: (1) It’s true that the participants sincerely want to
comply with law/are often risk averse, but the perverse thing is that
participation has to be combined with cynicism about tools/distribution: everybody
agrees to ignore the distribution. (2) As for the characterization of “periphery”:
The terminology makes sense to me because these issues are beyond c’s
boundaries: none of c’s business. Maybe there’s another discourse about
boundaries—this is an invasion. (3) One thing that’s valuable to the industry
is not to lose, ever, and that may explain some of the dynamics.  

Buccafusco: dividing the world into software and nonsoftware
might make more sense—allowing © in software is a key problem. Ripping DVDs is
at least plausibly in c’s wheelhouse.

A: yes, probably an original sin, but probably a bad idea to
design 1201 to protect distribution of video as well.

from Blogger http://tushnet.blogspot.com/2023/08/ipsc-breakout-session-4.html

Posted in Uncategorized | Tagged , , , , | Leave a comment

IPSC Breakout Session #3: Distributive Effects/Art and Music

Posted on August 3, 2023 by rtushnet

Bita Amani, Some More Equal Than Others: Critical Contexts
for the (False) Promises of Intellectual Property Rights       

IP and sustainable development goals: What is equality
itself? Often formal equality is presented in the pretext of equal application
as if IP were immune from feminist and critical race critiques. IP rights
define boundaries of exclusion, inclusion and belonging, tied to nation-state
building. Substantive equality is in many international instruments, and can be
found in the exceptions for the blind/print-impaired.

Alfred Yen, A Federal Rules of Evidence Perspective on
Forensic Musicology and Music Copyright Infringement

Juries hear a lot of forensic musicology evidence they
shouldn’t which confuses them and leads to poorer decisionmaking in music ©
infringement cases. Contrary to FRE, experts testify about subjects w/in their
competence that are irrelevant; experts testify about subjects that are
relevant, but beyond the expertise of musicology.

Why? Courts make errors about the law of infringement,
leading them to think that certain things are relevant when they are not.
Courts also have mistaken assumptions about the field of musicology; abdicate
their duties under Daubert/Kumho Tire.

Musicologists in litigation: (1) identify formal
similarities—instrumentation, chord progression; (2) opine on how aesthetically
similar/significant those similarities are; (3) opine about the rarity of
similar features; (4) opine that copying did or didn’t occur.

The case for forensic musicology: expert music testimony
bears resemblance to handwriting analysis; they can both identify and evaluate
similarities. At some point, the similarity is enough to say it’s the same
source—same writer or same work. We assume an expert can better ID similarities
that prove or disprove the proposition in question.

Legal mistake: collapsing copying and improper
appropriation, primarily through imprecise use of substantial similarity. Same
concept of similarity governs both elements of the copyright claim! When the
expert witnesses testifies that the two works are aesthetically similar, that’s
not relevant to copying—not every piece of aesthetic similarity is probative of
copying—same instrumentation of drums and electric guitar for a rock song. But
once this happens, the jury starts to think that, if they sound a lot alike,
they must be copying. Jury misled into thinking that aesthetic similarity, not
probative similarity, establishes copying.

Mistaken faith in musicology: let’s find an expert to talk
about copying. FRE 702: testimony must be product of reliable principles and
methods and expert reliably applied methods to facts of case. Musicology
articles talk about copying all the time. Typical “accusation” is that they
identify similarity. There is no method of convincing; it is pure intuition.
Not product of reliable principles or methods. Not reliably applied—what constitutes
proof is entirely up to the expert. Good musicologists do things like examine
handwritten manuscripts that show drafting, not listening to two different
pieces of music dozens of times.

Which similarities matter is up to the analyst and how many
similarities matter is up to the expert; this is the kind of ipse dixit that
district courts shouldn’t allow. Couldn’t test it if we wanted to. At least
w/handwriting we can test them on real exemplars. Do we include the chords or
not? Does it matter if similarity is only in theme and not in chorus?

Secondarily, what about rarity? When a musicologist says “I’m
not saying these are copied but these similarities are really rare,” if they’re
making an empirical claim, their foundation is shaky. [Chris Buccafusco and I
have a paper in progress about this.] You’d want a big sample or a representative
sample to make such a claim, but no one has such a sample. Second,
musicologists claim expertise not by looking at representative works but
unusual works—the leading examples in the field, not garden-variety examples. Defendant
can do this more concretely by showing common examples. P would need reasonable
claim to comprehensiveness or representativeness, subject to cross-examination.

Calculating probabilities is even shakier—notes aren’t
randomly distributed and the chance that one note follows another is not independent
of the note. Also, the fact that something is unlikely in any given instance
doesn’t mean it’s unlikely to have happened by now: law of large numbers. If
you’re one in a million, there are over 300 of you in the US, so the chances of
overlap are really quite large.

What’s left? Musicologists should be allowed to testify to
the existence of formal similarities, but not on aesthetic similarity or
significance; about rarity if they have a foundation; not about whether copying
did or didn’t occur.

Thus: here are similarities; here is some information about
how common those similarities are; now the jury decides. Could also bifurcate
into copying then improper appropriation.

What if anything does this say about sufficiency of
evidence/ability to survive sj?

Betsy Rosenblatt: hinging on experts is wealth-dependent; is
there any way to get rid of that reliance? Maybe special jury forms.

[in response to Q] If musicologists have this problem the
rest of us do too. This is often poor res ipsa analysis.

RT: bifurcation might not work well b/c copying drives so
much, but worth testing empirically.

[rarity: the inferences are unfounded—that is, what’s the
point of testifying about rarity? It’s either about copying, which they don’t
have the data for, or aesthetic similarity, which they shouldn’t be testifying
about.]

A: Evidence saying: A brick is not a wall. Stuff that isn’t
proof but that is relevant is permitted. Goes back and forth on whether rarity
is relevant to an overall determination even if the ultimate conclusion is
beyond the expert. That’s why he’s interested in the expert’s foundation.

Kristelia Garcia: the bullets you give apply to the
plaintiff’s side. But defendants’ experts are doing something completely
different—identifying prior art, which cuts off the need to identify whether
there was copying. This can lead to settlement/end of cases: Even if there was
copying, we don’t know they copied from you! The musicologists are now
completely split into P and D sides.

A: that’s what tells me they are unreliable! But prior
examples do at least provide a foundation for testimony. But do you need an
expert to infer lack of copying if it’s already out there ten times?—either the
plaintiff doesn’t own it or the defendant didn’t copy it from the plaintiff or
it increases the inference that it was coincidence.

Garcia: most D expert reports don’t opine on whether there
was copying, just on existence of prior art.

A: not as critical of that; allowing dodgy evidence to push
a case forward is his target.

Amanda Levendowski, Open Source Perfume         

Not a conventional copyright negative spaces story, but not
a proprietary space either. Perfume is exclusive and exclusionary; perfumers
are more scarce by some counts than astronauts; most are white men who trained
in a small town in France.

Primer: Le Labo is a “mainstream perfumery,” not niche;
owned by licensees Estee Lauder, designed by outside “nose,” who works for a “composition
house” that develops perfumes for other people. Santal 33: Composed of “notes”
with top, middle, and base, with an “accord” of sandalwood etc. notes to create
a specific type of scent. Fragrance family: woody.

Proprietary perfume: Iconic hues are functional, so
unprotectable. Method and “captive” molecules are patented by composition
houses, but rarely by perfumeries. Fragrances aren’t fixed, which undercuts
copyrightability—Chanel ad “the fragrance becomes you.” Trade secrets can’t
deter competitors or composition houses w/GC/MS machines that can discover the formula,
but they prevent aspiring perfumers who don’t have the machines from sampling
scents—they’re only available in large quantities and they are often not sold
outside of commercial contexts.

Shalimar: white guy from France who never visited India creates
a fragrance inspired by an Indian ruler who loved his wife. Colonialism/exoticization—the
fragrance family “Oriental” was coined by the nose to describe the family of
which Shalimar is a part, and it’s been dropped by the house but still in use
elsewhere despite protests.

Perfumery can be democratized, though. Open Source Smell
Culture: perfumers can share formulas using CC licenses. CC has limited impact
on formulas, but applies to evocative descriptions and destroys trade secrecy.
There are ways to make these more shareable/accessible to aspiring perfumers.

Glynn Lunney: why use an existing open source structure?

A: resource constraints—but open source hardware only has a set
number of licenses, and probably won’t create one for perfume. Thinks that
perfumers can get by with existing CC licenses; tailoring a perfume-specific
license could come later.

Q: does Le Labo go after dupes/smell-alikes?

A: Dupe houses exist—a lot of protectability comes from
trade dress/branding/packaging, not from the scent. Dossier is a classic dupe
provider. Houses don’t like when you borrow an accord; that will get more
attention than a straight dupe.

Q: you can rent a GC/MS machine.

A: but aspiring perfumers don’t know that “shooting the
juice” is even a thing. Until you get deeper into subreddits of perfumery, you
won’t find out these machines give you a formula. And even if you get a formula
to work with, it gets hard/expensive.

Q: suggestion you describe the people you are discussing in
more detail—not familiar with chemistry until they get there, often marginalized.

Emma Perot, Music Copyright Ownership: Factors Behind the
Surge in Writer Credit and Rights Clearance

Why so many writers on songs? Why so many interpolation
credits? Law is part of it, but industry factors also affect this. Blurred Lines
had a big impact in legal circles and in public opinion. Other big cases: Led
Zeppelin, Katy Perry, Ed Sheeran. Before Blurred Lines, disputes over sampling:
Bridgeport v. 9th Circuit; standard for originality is contested; subconscious
copying accepted; access in the age of the internet is tricky.

Still, high standard for striking similarity, and inverse
ratio overruled in the Led Zeppelin case.

Risk mitigation: add writers preemptively to avoid a
problem; clear samples; add writers where accused; add writers to protect
reputation whether the legal standards for infringement are met or not.

Practical influences: “change a word, take a third”; producers
now part of songwriting process; songwriting camps; files shared digitally can
be altered by many; sampling/interpolation as common techniques.

Change a word, take a third: producers/managers/artists may
take credit to get revenue. Big artists ask for 20-30% credit even when not
present for composition. Producers are in the room too, creating music and
beats, b/c of how tech and creative processes have changed. Common in hip hop
and R&B but now the norm in pop as well. Producers now may make their own
songs, eg DJ Khaled.

Songwriting camps: labels bring people together for a
specific project. Rihanna’s Rated R album: over a dozen people gathered in
California for 2 weeks. Inherently leads to many hands being involved. Norm in Nashville
but more recent in pop industry. Used in Kpop—network of 700 songwriters. Need
new material very quickly.

Files shared digitally, encourages more editing; their
contributions are credited.

Sampling: labels encourage use of existing tracks; listening
audience likes nostalgia; past hits are a recipe for success; catering to
social media and music streaming algorithms like TikTok charts; remakes of relatively
recent songs are popular.

Dynamics are risk mitigation—do not lose money—and practical
revenue maximization—make more money.

Tang: is this more songwriters per song or pop song?

A: Billboard top 100, not only pop but that tends to be the
source.

Q: can you look at whether rise in hip-hop/R&B accounts
for increased number of writers? More collaborative.

Q: Big artists demanding songwriting credit—songwriters get
so much less from streaming. Why does Beyonce care about songwriting when it’s
so much less than what you get from the recording on streaming?

A: may be that streaming is making less than normal analog
sales. Artists still like to be known as songwriters b/c it adds to their
reputation.

RT: Labels may see this as getting something for free b/c
the songwriters’ share is set by statute/agreement, so they can readily give
away money that would otherwise go to a smaller number of songwriters—are there
any costs to the labels? Even when they front the costs, do they deduct it from
the artist’s revenues?

A: yes.

Rosenblatt: a lot of self-dealing—licensing samples from
yourself so you can pay yourself. Real question: why aren’t we seeing WFH? [I
thought we were with Spotify, at least.]

Newman: coauthorship isn’t purely a matter of contract,
though it’s true that we rarely second-guess an agreement. But it sounds like
the industry has disregarded the theoretical rule.

A: songwriters have tried to push back, but not with
success. But there are questions with reversion rights. Can the real
songwriters terminate?

Chien-Chih (Jesse) Lu, Determining Music Copyright
Infringement in the Taiwan Context

Sleeping Beauty case: P performed song at school in 2001; D,
also a student there, composed a melody 90% the same. District court found that
access had not been proven. Did not adopt rule of striking similarity; did not
accept “reasonable probability of access” as relevant standard. The district
court wanted a particular chain b/t the P’s work and “wide dissemination.” But
there was no evidence that the work was in an archive or library. Although the
performance was recorded on CD, and became a sensation after the end of
semester performance, the district court assumed he work had not been widely
disseminated. Needs more flexibility, where there might be no records.

Liangshan Love Song Case: local tourism office invited P and
D to compose; D took P’s version and changed more; differences in composition
didn’t affect similarity of main melody. But we need to figure out who is the
original author. Without a registration system, it can be difficult to identify
who came first.

Taiwan’s copyright law has changed over time for the same reasons
US did. Developing country shifting from farming to global markets,
manufacturing and tech. Adopted IP rules to fit into global system. Criminal IP
enforcement was very strict in order to fight image of “pirate kingdom.”

Taiwan should adopt the idea of “striking similarity.” Use a
reasonable standard to demonstrate what is highly suspicious. If the degree of
similarity is not high, more substantial evidence of access should be required.
But if the degree of similarity is high, only reasonable opportunities should
be required. Judges sometimes hire experts in Taiwan, could be used for substantial
similarity. Amicus briefs can also assist the court.

Another issue: not a focus on actual harm. Instead of civil
litigation, Taiwan copyright owners tend to raise criminal cases to get
stronger penalties/save money. High cost to public interest b/c criminal
prosecution is carried out by prosecutors and increases judicial workload.

Last issue: Taiwan has no registration system; registration
could help with information problems like questions of authorship and
ownership. Gov’t is interested in improving licensing, but focused on TM and
patent. Ministry of Culture might be better home for ©.

RT: I worry about infringement creep—we have seen in the US claims
of striking similarity of 4 note sequences, which is very different than 90%
overlap. Maybe striking similarity is acceptable for 90% overlap, but it is
very risky to use a sliding scale without having a very strong lower bound for
what is potentially actionable.

from Blogger http://tushnet.blogspot.com/2023/08/ipsc-breakout-session-3-distributive.html

Posted in Uncategorized | Tagged , , | Leave a comment

IPSC Breakout Session #2: mostly copyright and then marijuana innovation

Posted on August 3, 2023 by rtushnet

Michael Carroll (w/ Peter Jaszi), Fair Use After Google and
Warhol

Codification is a big deal; clarifies that fair use is a
distinct doctrine, whereas well into 20th century courts were using it as
noninfringement. Courts weren’t using four factors before that. And it creates
statutory construction issues that justify SCt intervention, which hadn’t
happened before.

Early stage: focus on fourth factor—fiscalization—Sony
arises as secondary issue/throwaway piece in context. Harper & Row gets
litigated as a First Amendment/free press case by Floyd Abrams; Court relies on
Nimmer. Campbell is the first case that squarely presents §107 construction
without other distractions, and the Court takes a fresh look.

Warhol reinforces transformative uses; creates a straw man
version where new meaning or message is enough and rejects that, but he didn’t
see that in the lower courts anyway. The movie adaptation of the book isn’t
inherently fair use. Likely to see the vocabulary of whether the secondary use
is for a “distinct purpose” increase in future cases, but functionally, it’s
the same inquiry as in previous cases. Court chose to rule quite narrowly in
part b/c of concessions during litigation. Second Circuit had buyer’s remorse
about Prince v. Cariou, but by only affirming the judgment, Prince v. Cariou is
still arguably good law w/the exception of the relevance of artist’s intent.
Emphasis on narrowness of opinion—the only use we’re ruling on—suggests that we
need to take them seriously, and the concurrence wants to drive that point
home. So he doesn’t think this undoes the progress of Campbell and Google v.
Oracle.

By defining purpose only as licensing image of Prince—use in
a well-structured conventionalized market—the majority gets to be dismissive of
the dissent; it doesn’t matter how much Warhol contributed b/c he created a
substantially similar image which is now competing with her photograph.

The analysis does raise the level of generality issue: how
do we decide what the purpose is? The dissent is very concerned about limiting
available distinct purposes—Whelan v. Jaslow distinction between
idea/expression. But he doesn’t see that in majority and its emphasis on the
specificity of the use. The majority says that limiting doctrines account for
Kagan’s examples—which involved use of expression—as well as the dissent’s own
copying and the Court’s too.

Commercial use in well-structured conventionalized market in
which the © owner’s work participates, then the first factor will weigh against
the use absent some additional justification. Google added interesting things
under the fourth factor. Purpose is now to be determined by objective evidence;
Court says nothing about what that might be.

RT: Does the Wind Done Gone have a different purpose from
Gone with the Wind? It participates in the same conventionalized market.

A: b/c it’s not a substitute—it appeals to a different
market.

RT: but no one would read WDG w/o having read GWTW. And we
could make up a reason why they’re substitutes the same as the Court made up
the idea that one would choose between the Warhol and the Goldsmith photo.

A: it’s a different experience.

Matt Sag: Why do you think the holding has to do w/markets?
Court also says: Difference can’t be a matter of degree, can’t be new
aesthetic—he sees that as explicit rejection of Cariou. We don’t even get into
the market stuff.

A: but the holding is specific that this is only about
licensing in commercial markets.

Rosenblatt: does factor four do anything that factor one
doesn’t do under this analysis? Your answer was that the two works didn’t
compete.

A: to the extent you think purpose still counts as
transformative, it’s a different user experience than the original.

Jessica Silbey: same work in different uses will sometimes
be fair and sometimes not?

A: yes. They chose not to say the making was not a fair use,
hanging on wall, and use in art history book—seem to be conceded fair uses at
oral argument.

Silbey: what about the language of the case about the
necessity of the use? Seem to demand need to take Goldsmith’s photo as opposed
to any other photo. That goes to why Warhol took it not why Vanity Fair took
it. You’re reading rationality into the opinion where there’s a lot of
inconsistency. Justifying the use goes to the commentary aspect.

A: b/c he’s not commenting on her photo, his justification
for using her photo is lower.

Hughes: Why is GvO broad and Warhol narrow? Neither case
says it’s broad or narrow.

A: scope of QP, since Warhol is only factor one and decided
as only licensing uses.

Zahr Said: what do we do about risk aversion as a result?

A: on the ground, have to give individualized advice and
write up guidelines that make similar arguments.

Fred Yen: Wish the Ct had been honest that it thought Warhol
was garbage and not transformative. All the other bits are disguising the Ct’s
appraisal of the value of the Warhol work, the very thing it says no one else
is supposed to do.

A: we have to give meaning to the verbiage.

Yen: no we don’t, not as scholars. Don’t say the text of
this opinion reveals the truth of fair use.

William Henslee, The Transformation of Transformativeness:
The Implications of Fair Use on AI

Fair use should be limited to uses in the preamble of 107;
has become a catchall for ripping people off. Sony v. Universal was wrongly decided
and there should have been a blank tape royalty. Campbell was transformative as
a parody. But using 100% of another’s work shouldn’t be allowed; transformative
became the main event. But any lawyer can generate a reason that there is a
transformation. Must add something new to the original work. The fourth factor
is about economics; the first factor should be about artistic expression. Warhol
licensed the photo b/c he knew that if he wanted to use another’s work he
should pay for it. [This is a factual misdescription—Vanity Fair licensed the
photo and sent it over; the public record does not contain details about what
Warhol knew/believed.]

Recently, lawyer got in trouble for submitting AI-generated
legal argument. But data mining has been found to be a fair use. Data mining
legal documents, work product, court cases, and then create work based on
previous work. The input was copyrightable; the output is not. Taken
copyright-protected material and transformed it into public domain
material—dangerous for law firms.

Copyright for humans only? We do have WFH, and could say
that someone does own the AI output as WFH. Other solutions: statutory license
for sampling. Bridgeport is a great case; 9th Circuit was wrong. Fair use is a
failure b/c the only way to know if something is fair or not is to litigate to
the Supreme Court, and you might not know even then. If we’re going to have ©
for computer code, we could have a different fair use rule for code.

Bob Brauneis: the preamble says “such as,” and it doesn’t
have a unifying principle—teaching, including multiple copies for classroom
use, is very different from many of the others. Hard to get a narrow rule out
of that.

A: most of those categories are nonprofit. Parody should be
on the list, as comment and criticism, and some commerciality is ok.

Chris Newman, Fair Use: Against Weighing

Balancing rhetoric is mostly not useful; use factors to
describe outcome that they were already heading to by intuition, or they
gravitate to one part that they treat as dispositive. Not everything has to be
transformative to be a fair use; we’d be better off focusing on why that is
than focusing on rhetorical sleights of hand. Instead, we should think of
factors as lenses for triangulating on same object—they’re interrelated aspects
of the same question. We should never find factors pointing in different
directions; they should be reconciled not outweighed.

Breyer in GvO is model: discussion of nature of copyrighted
work should provide a qualitative grounding for the analysis and for
distinctive lines of precedent for different types of work.

You have to know what you’re looking for and we don’t have a
clear articulation of what we want from fair use. Wants to give mid-level
frameworks to guide analysis but that aren’t so restraining that they can’t
evolve and cover new situations. There are actually two different paradigms of
fair use and we should analyze them separately, even when one case raises both.

What are we looking for? “Supersedes the objects” of the
original = not fair use. Two dimensions: internal, expressive objects of the
work—what does it do for the person who consumes it? Folsom paradigm.
Originally this was internal to substantial similarity, and it still is and
should be until we get a better notion of substantial similarity—claim
construction/scope/boundary—charting boundary b/t derivative works and things
changed enough that they aren’t w/in the scope of © owner’s rights. If there’s
enough change, it doesn’t matter if it’s commercial.

Separate set of problems: external, instrumental—Sony
paradigm. May be using same thing for same purposes but it can still be ok.
What is the © owner justly entitled to and to what extent are you interfering
with it? It’s a nuisance analysis. Nonexpressive use and market failure are
part of this analysis.

Folsom: Story is clearly just focusing on the idea that,
sometimes, infringement is prima facie quite difficult b/c the boundaries
aren’t clear. Transformative fair use ought to be limited to this inquiry: have
the original expressive purposes/concrete expression used to further those
purposes been altered beyond the proper scope of the © owner’s claim? Social
value isn’t relevant to Story’s analysis—it’s still the same work with the same
expressive purpose. If it is transformative, we don’t care if it’s commercial;
if it’s not transformative, we care about social value. Transformativeness
should ask whether the content changes serve some expressive purpose distinct
from the original or its derivatives and so the resulting work is no longer
serviceable as a substitute for original or its derivatives. Orange Prince
doesn’t satisfy this standard b/c it’s still a picture of Prince with a lot of
artistic meaning added (the artistic contribution has not been removed or
effaced) but Leslie Nielsen in the position of Demi Moore does b/c the picture
is no longer a picture of Demi Moore. Blanch v. Koons = collage creates a new
gestalt; but String of Puppies is just added color.

Warhol is a Folsom case: it’s the same artistic depiction,
altering but not transforming its visual impact.

In the Sony paradigm, you don’t have to be transformative.
Conflicts b/t instrumental means of using the same work to exploit value. Is
the value being exploited properly w/in owner’s use and enjoyment? Would a prohibition
merely destroy value w/o helping © owner exploit its own value? GvO is this
kind of nuisance dispute—it’s clear that the value they’re exploiting isn’t the
value of the expression of the code but the fact that people are locked in to
knowing this code. Using factor two allows us to do this.

Warhol: external, instrumental inquiry into use can focus on
the specific use being made; you can say that the art world uses are outside
the scope of what Goldsmith has the right to expect. What sorts of demand is
the author seeking to fulfill? Is the accused use fulfilling that sort of
demand?

Search engine cases: the initial copying is not
transformative but it’s nonexpressive, which isn’t dispositive but it creates
value that doesn’t satisfy demand for original. And the output is a Folsom
situation: the output doesn’t use 100% of the original/vastly reduced size or
snippets.

[hard to make this labeling move post Warhol, which leans
into transformative purpose like the Court’s own use of images and also insists
that the inquiry is into use and not work: for the latter, if seeking
consistency w/Warhol. perhaps you should also reframe the first part of the
inquiry to analyze the use and not the work?

Larger issue: You’ve just made the first inquiry depend on
one’s conception of the derivative work right, and if you do the second inquiry
in Warhol you also need a concept of the scope of the derivative works right.
Relatedly: Annie Liebovitz doesn’t own what Demi Moore looks like and does own
the specific expression that Naked Gun 33 1/3 copied, so taking away Moore’s
face shouldn’t affect whether the expression was copied.]

Betsy Rosenblatt, Considering the Role of Fairness in
Copyright Fair Use

Could we make fair use more fair by considering equity and
distributive justice and favor the historically disadvantaged/less privileged
in fair use inquiries? Just v. unjust enrichment. Fair use allows less privileged
groups to develop skills, support educational equity. The other side:
privileged players are more likely to prevail in fair use litigation and courts
are more likely to recognize transformative merit of famous or popular user.
Litigation is not wealth-neutral. Could allow/permit exploitation. A teenager
who sells a fanwork based on a tentpole movie feels different than a movie
studio that creates a tentpole movie from her short story. In the latter case,
they weren’t deprived of expected income, but it feels like unjust enrichment
in the moral sense. © can be conceived of as helping the less privileged.
Dilettantes can afford to create w/o ©, but © allows professionalization.

It would be doctrinally permissible to distinguish between
relative status and cultural market share, but extremely hard to implement. Economic
and sociocultural disadvantage, historical oppression put into the same bucket—they’re
very different in every other way but they’re barriers unrelated to inherent
ability to create works of authorship.

Complications: we’d have to think about when in someone’s
career we’re looking at. Time of alleged (initial) infringement v. time of suit.
Look through use and user to the beneficiary. HathiTrust was done by big
companies, but on behalf of less privileged users. We also have to look through
© owners to look at ultimate beneficiaries of ownership. This might be too hard
to implement. But we could include in jury instructions that we could consider
these things. Might prevent unnecessary litigation. Might give marginalized
creators more ownership (v where they don’t own any rights in an infringing derivative
work).

Ruth Okediji: what does the Court’s equal protection jurisprudence
have to say about this?

Jacob Victor: kind of a market failure argument! If we are
inquiring into ability to pay etc.

Kristelia Garcia: how do you deal with generative AI? Fan
fiction in the style of Betsy Rosenblatt: is that a small creator in need of a
leg up or a help for another small creator?

A: may depend on who ultimately benefits from the existence
of LLMs, which we don’t know the answer to yet. Fans have very mixed opinions
about having their works used, especially for attribution and the fact that the
works were created in a gift economy; a machine provides neither attribution
nor participation in the gift economy.

Lucy Xiaolu Wang (w/ Nathan W. Chan), Is Grass Greener in
the Gray Zone? Legalization and Innovation in the Cannabis Market

Most widely used globally, 18% of US population used in
2019, legal market globally estimated at $100 billion by 2026. Medical/adult
use legal in 50+ countries; still most widely trafficked and abused illicit
drug. US: federally illegal but patented by fed. But we have little evidence
for medical treatment.

Documenting trends of R&D in cannabis market, estimate
causal effects of legalization, and examine heterogeneity of policy
responses/mechanisms.

856 US registered cannabis trials, 559 in US. About 80% are
in a single states. Funding: 6% industry, 47% NIH. Looked at treatment of
conditions; usage/effects on body and function of cannabis receptors; abuse
factors, including risk factors for abuse.

Medical legalization doesn’t appear to have any impact,
while US trials increase mildly after adult/recreational use
legalization—consistent w/anecdotes on creative use of dispensaries/wider
access to more cannabis users for trial recruitment. Not enough medical
R&D. More patenting post recreational legalization, but mainly
downstream-oriented.

from Blogger http://tushnet.blogspot.com/2023/08/ipsc-breakout-session-2-mostly.html

Posted in Uncategorized | Tagged , , | Leave a comment

IPSC Breakout Session #1

Posted on August 3, 2023 by rtushnet

Too much to choose from; I prioritized things that were new
to me.

Mark Bartholomew, A Right to Be Left Dead

If we can create new works in a dead author’s style after
they’re gone and create conversations with them (including loved ones) after
gone, and populate new entertainment with dead actors, isn’t that something
new? NY State created a postmortem right of publicity specifically for computer-generated
likenesses; Louisiana did so too. IP Subcommittee on AI: was a lot about
whether we need a national ROP to deal with this. Wants to focus on the
relevance of death.

How do other areas of law treat relevance of death? Map onto
3 main rationales for ROP.

Dignitary interests: false light, IIED, privacy typically
expire w/person. There is some exception for actions immediately surrounding death
or moment of burial. Death closes things off.

Consumer protection: laws meant to protect consumers, death
doesn’t matter—Princess Diana’s image on collectible plates—TM claim was maintainable.
So too w/false advertising. Makes sense b/c consumer protection law’s goal is
information safety and not vindicating rights of decedent.

Property law: somewhere between all and nothing. Decedent’s
operative intent to extinguish property right—when you bury people w/jewelry by
their choice, no estate tax; it just disappears. Body parts. Forced share laws
for spouses. To fix suboptimal postmortem allocations.

Need and precedent: how is AI being used to reanimate
people? One use is grieving: used to create a chatbot w/specific person corresponding
to a past entity. Advertising: false endorsement potential. Actors’ strike is in
part about who controls past performances if they’re being used to populate new
media productions? Great deal of variability in postmortem ROP. Okla and Indiana
give 100 years, potentially infinite in Tennessee. Entertainment: stronger ground
to articulate right in performances; incentive story makes more sense for this
than for other invocations of right of identity. Stare decisis: can we really
fight this any more?  We can at least
fight how we structure the rules of postmortem ROP when it’s only found in
about half the states.

Solutions: better ways to articulate postmortem wishes,
including a way to abandon postmortem publicity rights. Putting it in a trust like
Robin Williams did might not be enough—might be fiduciary duty to use it
commercially.

Duration: goal is to recreate original; the right shouldn’t
extend long past death, because the key is generational change/asserting
difference from previous one is a key mechanism.

Prior exploitation requirement? Need to have used identity
commercially during life in order to trigger postmortem right. Death is a
demarcation point in many circumstances; we should have a mechanism for free
use for research and other purposes for most people who didn’t have opportunity
to commercially leverage persona during life.

Fred Yen: why analyze things through interests of dead
person at all? Couldn’t we look at it through the interests of the living? Grief
trolls—sending a person messages from deceased loved ones would be IIED; copyright
can handle performances, collective bargaining can handle digital models in
Hollywood. There’s lots of reasons to keep the dead from having too much
influence over our lives.

Mark McKenna: many recent issues over use of personality after
death would have been unanticipated during their lifetimes. How practical is it
to imagine needing affirmative steps to anticipate those uses? Why not just a
set of defaults?

[My thoughts: Intro sounded to me like classic anxieties
over what currently seems like realistic tech of representation—putting words
in mouth that never said happened in written word, in print, in image, in
cartoon, and there have been repeated moral panics about how everyone will
perceive these as real/deceptive; we just keep resetting our expectations.]

Consumer protection/TM: why would ordinary consumers care
about consent of estate or believe there was unified control? And why would
endorsement of estate equate to endorsement of celebrity in consumer’s mind?
Assumptions skipped over in TM/false advertising analysis.]

Lucas Osborn, Christianity & IP

Core: creation, Fall, redemption, restoration. Genesis: G-d
said, “Let us make man in our image, after our likeness. And let
them have dominion over … all the earth.” Humans are categorically
different from other animals and have some special relationship to G-d, our
creator; we have a purpose in the world. Thus, we are created with creative tendencies/desires
to create.

But all is not well because of the Fall. We also have a
tendency to sin. The legal construct we use for that is the wealth maximizer,
who maximizes their own happiness. We also sometimes act irrationally b/c our
reasoning is messed up. Nonetheless there is still goodness in the world. That
can map onto behavioral law & economics. We will not always act justly or
see correctly what the just thing is to do.

Humans will generally not create at maximum potential or
create things for the common good or use creation in a just way—incentives may
be able to intervene here. Example: drugs that are overpriced; but even within
Christian worldview, fallen humanity may require attending to incentives.

Deborah Gerhardt: Tree of knowledge is part of the Fall. Jewish
theology sees this as opportunity to do good, to choose good or evil, a
maturation of knowledge. We wouldn’t need rules if there was no difference.

Jeremy Sheff: Spread of knowledge—evangelical perspectives
may differ from nonevangelical. Catholic tradition about authority over congregation/knowledge
is very different. Control over text is part of some traditions, along with
control of knowledge that might conflict with or comment on the text. Can you
separate those issues from the theological issues? IP is as much about distribution
of knowledge as it is about creation.

Dustin Marlan, Trademark Disclosure

TM registration requires disclosures just as patent does:
undertheorized. First use/use in commerce date, drawing of mark, goods and
services, owner, domicile.  Why? Protecting
brand investment and consumers?

The TM bargain: registration can be made in a form prescribed
by the Director, so there’s some statutory flexibility if we think policy justifies
it. Specimens: one per class, as currently required.

Benefits: nationwide right of priority, evidence of
validity/exclusive ownership, right to police imports, incontestability after 5
years, symbolic benefits like ®. Constructive notice to the public is often
thought of as good, but it can put the claimant on the radar (as trade secrecy
is an alternative to patent)—people could find you and send you a C&D;
admissions could be detrimental in later LOC analysis, e.g. conceding to a disclaimer;
might want to file in a jurisdiction without disclosure like Trinidad &
Tobago to avoid tipping off the market; icky mark or brand could be unpalatable
to the public.

Could TM disclosures be tweaked to make them more
consumer-friendly? TESS is a mess. Could it be more consumer-friendly to allow
more comprehensibility to the lay public?

RT: not sure that disclaimers ever matter to future rights,
but collateral estoppel/B&B is a big deal now, as is the loss of potential
defenses after Jack Daniels. Also I’m not sure that you can ever make
registration an engine of consumer protection b/c people never look up classes,
first use date, etc.

Justin Hughes: analogy to patent applications—written description
and enablement are quite different. Classes/goods and services might be the closest
analogy to metes and bounds.

Mark McKenna: natural to think about goods & services as
description of claim in a world where the claim is for a word mark; the
boundaries were clearer for Nike. But many other applications are less clear
for what they are, and we need more rules around that. Courts don’t follow
PTO rules about dotted lines; they aren’t careful about what the mark consists
of or comprises, especially to get to inherent distinctiveness. That’s where
disclosure rules need the most attention: definition and relation to what
courts are going to do with the registration.

Alex Roberts: Different reasons to choose not to use system—stealth
registrations in T&T are different versus “we define trade dress when we
choose to sue over near copies”; Mattel enforces a lot against pink but doesn’t
register it so it doesn’t have to define a specific color.

Bartholomew: Registration is optional, not mandatory as with
patents; is this the right number of sweeteners to encourage people to register?

Hughes: could compare cases where they narrow the claim of
the TM registration v. narrow the claim of the patent—Maker’s Mark case where
they have a registration for a wax seal but claim only red.

McKenna: and the reverse, where they register narrowly but
claim broadly.

Matthew Sipe: connection here with Katyal’s project where
public doesn’t get everything it needs for full agency in consumption.

Elizabeth Townsend Gard: people don’t register color as trade
dress—interesting to consider why not. Same with word marks—they register
standard word mark when color is an important part of what they’re doing.

Lorelei Ritchie, is this generic 😦 ?

Consumer perceptions need evidence—surveys, declarations,
dictionaries. Consumer perceptions change over time. Can a generic term be
ungenericized? After Booking.com, maybe it can be because consumer perception
is key. Evidentiary standard is not standardized. Merely
descriptive=immediately conveys information about feature, function, or
characteristic. But generic? The thing being named, or if it’s a key aspect or
subgroup, which is where we start running into problems.

Possibilities: allow acquired distinctiveness for any
asserted mark; require acquired distinctiveness for any asserted mark
(requires legislative modification). Should TM registration be w/o much
examination? Allow people to file claims and then go from there.

Another: expand consideration of a “mixed record” where
there is some generic and some nongeneric use. Many people use “google” to mean
“use Google.” Could expand on that more.

Legal standard: preponderance in inter partes case; ex parte
move towards preponderance. She argues that they should go the other way: clear
and convincing for everything. Why? Because generic terms get no protection.
Should look for abandonment and fraud where we have a higher burden of proof.
Consider preclusion as well. W/genericness, we are talking about the same
thing, so preclusion would be appropriate.

Unprotectable generic terms should be precise, like apple for fruit (note that this is a subset). Not a key
aspect or subgroup. “Apple candy” is not generic for candy, it’s just a flavor/key
aspect. Should be descriptive.

Relevant consumer: perceptions can determine commercial
strength but the inputs are often the same as for conceptual strength—dictionaries
w/lower-case references. Keep up with the culture by considering social media. Evolving
norms of communication and branding both have no punctuation/capitalization:
lululemon.

Hughes: Agrees that generic means thing: “highly descriptive”
doctrine should be used to moderate this problem for things like apple candy.
Relevant consumers for determination of genericness should be different for
more narrow descriptiveness group of consumers. Singer was recaptured from
genericity, as was Goodyear (probably).

McKenna: disagree w/Hughes b/c categories aren’t found in
nature. Is diet soda a category or a subgroup? It’s market definition. Is apple
candy a thing? That has no abstract answer. In favor of calling more generic:
once you introduce secondary meaning, you’ve told defendants that the case will
cost hundreds of thousands or millions, whereas genericity is more easily
managed. Before Booking.com there was the competitive need test and that’s why
failure to function has had a resurgence.

Roberts: Nike just got SNKRS w/vowels removed—that’s the Booking
problem.

RT: [I agree on the subgroup issue: soup spoon; car is a
subset of vehicle; apple for fruit is a subgroup; key aspect may be another way
of saying market-defined subgroup.

Spectrum of descriptiveness/ “highly descriptive” category will
likely defeat your attempt unless you take more steps to put that off limits
too.

Clear and convincing: basis in statutory text? There’s a
common-law case for treating fraud differently but I don’t see that for
abandonment or genericity; the “harsh consequence” is the natural consequence
of choosing a generic term. (Not to mention that these days you can still get
unfair competition protection under §43(a), which is not the case for patents.)]

Deborah Gerhardt, Trade Dress Edges

Where does trade dress begin and end? The definition expanded
far beyond packaging/display. Color, product design. This matters b/c we’re supposed
to apply functionality for trade dress, and we often use Seabrook for
distinctiveness, whereas there aren’t many disclaimer practices for trade dress
and there are for word marks. So there are doctrinal differences.

Possibilities: (1) Trade dress a subset of TM (based on the statutory
language in 1127); (2) partially overlapping (Two Pesos, where text might be
included in some trade dress; INTA view of overall look and feel without
functionality); (3) Restatement view of distinct categories. But do you have to
apply functionality analysis if there’s a word involved?

PTO doesn’t use this categorization—it uses text, design,
sound, etc. Traditional/nontraditional is also unhelpful and does no doctrinal
work.

Tyler Ochoa: we call functionality genericity when it
applies to word marks. [Not sure that’s entirely true—the AOL “you’ve got mail”
case is functionality more than genericity.]

Trade dress got overexpansive; Two Pesos got cut back in
Wal-Mart in particular with the tertium quid versus product configuration/packaging.

A: in sound marks we do see genericity and functionality
merging—alarm sound.

RT: Consider substantial similarity—it’s the standard for
everything in © but we recognize that different genres may require different
analysis. Choral music requires experts to tell you what might count as similarity
of expression in choral music, etc. The value is in finding patterns and so I’m
comfortable with being in the subset box (1).

Betsy Rosenblatt: willing to go further—there are lots of
words that are informationally functional and should be denied registration for
that reason for relevant goods/services.

Roberts: what’s inside the “trade dress” circle? Are sounds and
smells trade dress, or are they something else?

[Things that can be inherently distinctive v. things that
can’t be as another important line?]

Sheff: identify where the distinction matters—functionality,
maybe; inherent distinctiveness not available for product design; practical
relevance: defining the unregistered trade dress is a big part of the job for
§43(a)—even though there may not be a doctrinal issue it is a very practical
one where we fight over how specifically we have to claim features.

A: Impetus: some sound marks are treated like text, and
others like product design—analytically unsatisfying. Rules for registering
text don’t work well.

from Blogger http://tushnet.blogspot.com/2023/08/ipsc-breakout-session-1.html

Posted in Uncategorized | Tagged , , , | Leave a comment

IPSC Opening Plenary Session

Posted on August 3, 2023 by rtushnet

 Matthew Sag, Copyright Safety for Generative AI

Not addressing whether training is always fair use in every
circumstance; explain how generative AI fits w/in existing law (nonexpressive
uses) and identify best practices to make generative AI fairer.

Non expressive uses: reverse engineering, anti-plagiarism,
search plus text data mining, snippets. Although Authors Guild didn’t discuss
it, that was very much about machine learning—the endgame was more sophisticated
analysis using machine learning, not just frequency tables. His theory: rights
of © are defined w/r/t communication of original expression to the public.
Nonexpressive uses don’t involve human enjoyment of the original expression
contained in a work and don’t lead to other humans potentially enjoying that
original expression. Thus, nonexpressive uses don’t interfere w/the kind of
interest © is designed to protect, which is why they are such strong candidates
for fair use.

But is generative AI really the same? Are the outputs copies
of the training data? Copying takes place prior to training; converted into
tokens and training is a process of adjusting weights in the model, not copying
tokens. Info acquired is much more abstract than the training data. The output
is a novel synthesis, not just a cut and paste, usually combines abstract
latent features learned from training data.

But CS literature on extraction attacks exists: in
successful attacks, models may be trained on many duplicates of the same work;
images associated with unique text descriptions; ratio of model size to
training data is relatively large. Generally the extraction attacks fail; only
succeed in edge cases. These are real risks, but rare. More common: The Snoopy
problem: text to image models are set up to learn attributes of ©able
characters: b/c same text description paired w/relatively simple images that
vary only slightly. Not limited to characters; other images repeated w/ minor
variations and consistently tagged w/same keywords: Banksy’s girl holding a red
balloon.

Implications: too much memorization undermines arguments in
favor of fair use. Hard to argue it’s highly transformative if training data
come out as output. May also have implications under fourth factor. What to do?
Ten suggestions in paper, but key: (1) reasonable measures to deduplicate; (2)
reasonable measures in training and deployment to reduce probability of
infringement—curate/pre-processing training data; using reinforcement learning
through human feedback; installing restrictions on model outputs. (3) reasonable
measures to safeguard privacy interests. (4) detailed records of works used and
their provenance.

Sonia Katyal, A Trademark Theory of Rebranding

Lots of rebranding out there. Cultural and economic factors
(change in what the business does over time, including startup drift and
embrace then rejection of crypto). Practices are undertheorized and
underregulated. Not transparent/information asymmetry for consumers. Internal
contradiction in TM: rebranding could be characterized as contradiction b/t law
and marketplace: marketplace gives incentives to rebrand, given lack of legal
regulation thereof. A single user can obfuscate its past identity from its
present. Source concealment/distortion.

Most common reason: change of ownership or structure,
merger, acquisition, sponsorship change, shift from private to public, demerger

Change in corporate strategy—globalization or localization,
diversify or divest. Change in competitive position or external environment
(attracts the most commentary). TM doctrine assumes marks are chosen for
potential for perpetuity. Allows tacking of priority w/continuing commercial
impression. Material alteration requires new filing. This rule facilitates
information asymmetry that can harm consumers by adding information costs. Distortion:
expansion of boundaries by continuing commercial impression (actual TM
boundaries). Also identity concealment—Worldcom to MCI; organizational
misdescription—renaming, reorganization, shell companies; source obfuscation, where
FB rebrands to Meta as distraction; affiliative disinformation—broader rebranding
associated w social movements.

TM law should facilitate disclosure or it will fall behind
market.

Sheff: What about licensing/merchandising in this system? That
deprives consumers of info in the way that you’re talking about; the law used
to treat it as a kind of deception but now we all love it.

Sag: Why isn’t this like bankruptcy? Fresh start.

A: good for companies, not so good for consumers.

Alex Roberts, Multi-Level Lies: Distinctive feature is that
laypeople are making advertising claims to friends, social media followers,
acquaintances, etc. Sellers make money by selling product but also by recruiting
others. Exploitative: 99% make no profit/lose money; stuck w/unwanted
inventory; exploitative based on false promises of wealth and miracle products—the
seller gets free advertising. Not surprising that sellers make false claims—they’re
desperate to sell. Lack knowledge/training; rules aren’t intuitive, including
risks of sharing true but unrepresentative experience. Nondisparagement clauses,
some of which purport to bind family members. Indemnification clauses (likely
uncollectable, but still scary). Many contracts—77%–say sellers can only use
marketing claims or materials provided by company; 71% say other claims must be
pre-approved (that seems contradictory). 57% say sellers could be liable for
false etc. claims, but only 24% reference or explain specific rules.

Many actual sellers (convenience sample) make effectiveness
claims—80% almost always/half the time; 85% make testimonial claims; 73% make
data-based claims (lose 1.8 pounds/week); 86% make positive subjective claims;
78% make business opportunity claims. 59% were aware that laws/rules constrain the
claims they make, mostly made aware by training or other reps from company. Did
they influence how you advertise? Many people said yes, but many people skipped
the respondents—47% said they almost always follow the rules. Many expressed
uncertainty that the rules applied to them. 46% said desire to sell outweighed
desire to comply; didn’t expect enforcement. 37% said the company didn’t say anything
about restrictions on claims or said that there were no restrictions.

Takeaway: companies, not lay sellers, must be held liable
for false, misleading, or noncompliant claims disseminated by sellers. FTC and FDA
think this already!

More enforcement; more training for MLMs with interactive onboarding/continuing
education. Expand social media monitoring.

Contract issues: limiting arbitration and class action
waivers would help; classify sellers as employees? Platforms should enforce own
TOS against these claims.

Katyal: are the influencer rules different/should this be
folded in?

A: biggest difference is that influencers get paid by the
seller and MLM sellers don’t. Affiliate marketers get a cut of sales; that’s
similar to how MLM works, so it’s pretty easy to see how they fit in, but you don’t
really need that piece because FTC has said elsewhere that the companies are
liable for claims made by sellers—the case law doesn’t require the endorsement
piece to hold the company liable for the claims.

Q: endgame?

A: even with the ones that aren’t snake oil, and many are,
there are risks of false income claims that should be guarded against. FTC
could go harder against supplement market in general; FTC/FDA should team up
more often.

Christopher S. Yoo, Common Carriage and Social Media

Hating on social media is bipartisan. Turning to common
carriage analysis. Declare as matter of statutory announcement that they’re
common carriers and hope that makes the regulation constitutional, as in Fl and
Tex. Can that be right? No. Turner Broad. Sys. v. FCC says that the Court has
to assess those judgments. Thomas once said that labeling a regulation a common
carrier scheme “has no real First Amendment consequences.”

What about common law? Rationales are almost all debunked
but one works.

“Affected with a public interest.” That’s not enough. It’s
an empty category. Newspapers were critical for disseminating information, but
that doesn’t deprive them of 1A protection for editorial judgments.

“Monopoly power.” That’s historically not been true. The
root is not in 17th century English common law, but 20th century laws creating
public utilities. Even if that were the case, should it make a difference to
the 1A analysis? Not in Tornillo, even though most cities were one-newspaper
towns at the time.

“Transportation/communication.” This is a useless category.
Lumping things together doesn’t teach anything (Oliver Wendell Holmes). Too
general to be helpful. There are tons of communications media that aren’t
regarded as common carriers.

“Quid pro quo.” Section 230 as a quid pro quo. But not clear
that common carriers are the beneficiaries of 230. Another statute enacted at the
same time says “nothing in this section shall be construed to create
interactive computer services as common carriers or telecom carriers.” There’s
clearly some immunity for entities that aren’t common carriers.

“Holding oneself out as serving the entire public.” It’s an
antifraud test, but easily evaded. DC Circuit says that FB, Google, etc. don’t
hold themselves out as affording neutral, indiscriminate access and that net
neutrality didn’t apply to any services involving editorial judgment.

The real motivation: syllogism, not 1A analysis.

Major premise: old regimes like common carriage must be
constitutional.

Minor: new regulations of social media are like common
carriage.

Conclusion: new regulations ok.

Major premise is problematic; SCt has never resolved.
Precedents uphold access regulation only for media in which one actor exercises
physical, not economic, control that social media lack. Invoking common
carriage accomplishes nothing/serves as a distraction.

Lemley: What do you do with telephones? We clearly apply
that to telephones.

A: dial-a-porn cases can help answer. Both sets of cases say
that even a common carrier offering services, you can also offer video over
which you exercise complete discretion. You can be a common carrier for only part
of your services.

Lemley: but only b/c we put telephony and cable in different
buckets. Why not exercise discretion over how and under what circumstances we
deliver telephone calls?

A: is the logical implication that there are limits on that?
Yes.

Interestingly, neither Fl nor TX bars discrimination in the
ordinary sense—deplatforming and shadowbanning are not the same things. Not a
true common carrier regulation.

Chris Newman: nondiscrimination in common carriage is about
passage, not behavior on the vehicle/harassing other passengers.

A: agreed. The author of the 5th Circuit opinion said we
reject constitutional challenges to nondiscrimination laws, but that’s about
economic discrimination, not speech discrimination. That’s handwaving.

Nicholson Price, Empirical Studies of Medical AI Patents (w/
Mateo Aboy & Seth Raker)

Applications are up even after Bilski, etc. Grant rates
dipped but then went back up. 3 of top 5 patent applicants are the 3 biggest
MRI makers. Disclosure is only ok; universities do better. Predictions to
contrary underestimated the cleverness of patent lawyers. We’ll get lots of patents
anyway; people figure out ways over around and through doctrinal barriers.

from Blogger http://tushnet.blogspot.com/2023/08/ipsc-opening-plenary-session.html

Posted in Uncategorized | Tagged , , , , , , , | Leave a comment

poop bag seller’s “compostable” claims are plausibly … false

Posted on August 1, 2023 by rtushnet

Natale v. 9199-4467 Quebec Inc., 2023 WL 4850531, No. 21-CV-6775
(JS)(SIL) (E.D.N.Y. Jul. 28, 2023)

Many lawsuits against “recyclable” claims have been
dismissed, but not this “compostable” one. Defendant sells “Earth Rated
Certified Compostable Poop Bags.” “On the packaging [of] the 60-count version
of the Product, Defendant represents that the Product is ‘Certified
Compostable.’ ” The packaging for the 105-count and the 225-count versions of
the Product are “identical in all respects except for the size” and likewise
state that they are “Certified Compostable.” Defendant’s website says that its
“certified compostable bags meet the ASTM D6400 standard for municipal
composting as well as the EN13432 Home and Industrial standards for
compostability.” It further says that the Product “is ‘certified for home
composting’ and is compostable at a ‘city compost’ facility.”

But: “On the back of the 60-count version of the Product,”
the packaging “includes a small print disclaimer in small font” stating:
“COMPOSTABLE IN INDUSTRIAL FACILITIES[.] Check locally, as these do not exist
in many communities.” The packaging also states “[n]ot suitable for backyard
composting.” A small print disclaimer on the side panel of the 60-count version
also says: “[s]hould only be disposed of in commercial composting facilities
where pet waste is accepted.” The 105-count and 225-count versions of the
Product also include small-print disclaimers stating: “Should only be disposed
of in commercial composting facilities where pet waste is accepted. These
facilities may not exist in your area. If you want to compost your pet waste in
a home compost, please ensure to only use the resulting compost on non-food
crops.” The disclaimers on the website require clicking a “Read more” tab; FAQs
don’t use relevant titles. Users must navigate to an innocuously titled FAQ:
“[h]ow should I dispose of the compostable poop bags” to access the
disclaimers. Plaintiffs alleged that “[n]o reasonable consumer would expect
that by clicking on ambiguously named or indiscriminate links, they would find
small print language on Defendant’s website … that would contain language
inconsistent with the representation that the Product is capable of being
composted.”

The FTC “has stated that ‘compostable’ claims on dog waste
products are ‘generally untrue,’ ” given that “dog waste cannot be composted
because it can contain harmful contaminants (e.g., E. Coli).” The EPA has
stated that “[e]ven in backyard composting … dog waste can contain harmful
parasites, bacteria, viruses, or pathogens.” Plaintiffs further alleged that
the phrase “if [the] city’s municipal composting accepts pet waste” was false
and misleading because “industrial composting of dog waste is not available in
the United States.” Plaintiffs alleged that “[n]o reasonable consumer would
expect that the small print language on the back and side panels of the Product
would contain language inconsistent with the representation that the Product is
capable of being composted” and that no reasonable consumer would “expect that
a ‘certified compostable’ dog waste bag would not be capable of being
composted.”

The FTC’s Green Guides use two relevant examples:

Example 2:

A garden center sells grass
clipping bags labelled as ‘Compostable in California Municipal Yard Trimmings
Composting Facilities.’ When the bags break down, however, they release toxins
into the compost. The claim is deceptive if the presence of these toxins
prevents the compost from being usable.

Example 4:

Nationally marketed lawn and leaf
bags state ‘compostable’ on each bag. The bags also feature text disclosing
that the bag is not designed for use in home compost piles. Yard trimmings
programs in many communities compost these bags, but such programs are not
available to a substantial majority of consumers or communities where the bag
is sold. The claim is deceptive because it likely conveys that composting
facilities are available to a substantial majority of consumers or communities.

The 60-count version sells approximately $0.15 per bag. Defendant’s
similar 120-bag product that is not certified compostable sells for approximately
$0.06 per bag.

Unsurprisingly, plaintiffs properly alleged economic injury
sufficient for standing. And they adequately pled deceptive acts/false advertising
under N.Y. G.B.L. Sections 349 & 350. They plausibly alleged lack of compliance
with the Green Guides, getting out of NY’s safe harbor provisions for when an
“act or practice is … subject to and complies with the rules and regulations
of, and the statutes administered by, the [FTC] or any other official
department, division, commission or agency of the United States.”

It was plausible that the disclaimers weren’t clearly and
prominently displayed both because they were on the side/rear and because they
contradicted the prominent “Certified Compostable” claim on the front, as well
as defendant’s admission in one place on its site that “industrial composting
of dog waste is not available in the United States.” There were further contradictions
in that the 60-count package said it wasn’t is not suitable for home composting
but the larger count versions state they are compostable at home, but not for
food crops. Regardless, plaintiffs also alleged that pet waste is not safe to
compost at all due to its release of harmful bacteria as it decomposes. This
was sufficient to survive a motion to dismiss.
“[W]here appropriate facilities are not merely limited, but are in fact
non-existent, the suggested language that customers should ‘check locally’ because
such facilities ‘do not exist in many communities,’ or that facilities ‘may not
exist in [the customer’s] area,’ is plausibly deceptive in that it suggest that
if consumers were to look for an appropriate facility they may find one, when,
in actuality, no such facilities exist.” This problem was compounded by the contradictory
disclaimers; if it wasn’t safe for backyard composting or eligible for
industrial composting, then it was plausibly deceptive because it wasn’t
compostable at all. The package didn’t explain what the ASTM standard meant or
why it would matter if the consumer couldn’t actually compost the product.

Warranty, fraud, and negligent misrepresentation claims also
survived.

from Blogger http://tushnet.blogspot.com/2023/08/poop-bag-sellers-compostable-claims-are.html

Posted in Uncategorized | Tagged , , | Leave a comment

allowing free “For Sale By Owner” ads doesn’t imply no agents will be involved

Posted on August 1, 2023 by rtushnet

Picket Fence Preview, Inc. v. Zillow, Inc., 2023 WL 4852971,
No. 22-2066-cv (2d Cir. Jul. 31, 2023)

District court opinion discussed
(second) here.  The district court
dismissed Picket Fence’s false advertising/unfair competition claims against
Zillow under the Vermont Consumer Protection Act and Lanham Act, and the court
of appeals affirmed.

Picket Fence is a For-Sale-By-Owner (FSBO) “publication
business” that permits private homeowners to list available properties directly
to potential buyers in exchange for a fee. Zillow provides an “online portal”
for the advertisement of property and realtor services to “the general public.”

Unlike Picket Fence, however, Zillow advertises that it
permits FSBO property listings on its website for free. Picket Fence alleged
that these claims were false. The problem was that, after a seller would post a
FSBO listing, Zillow would allegedly “divert potential buyers to its paying Premier
Agents” by “stripp[ing] out all contact information” for a FSBO listing’s owner
or by posting the FSBO owner’s contact information beneath an advertisement for
a Premier Agent’s services, inducing potential buyers into paying additional
charges. This contradicted the FSBO promise because a FSBO advertisement was
allegedly “one that allows a person to advertise their property so that
potential buyers can see the advertisement and contact the owner/seller
directly without the use of a third party intermediary.” By making Premier
Agents a part of the transaction, Picket Fence alleged, Zillow was falsely
advertising that it permitted FSBO listings for free.

The deception was allegedly that a FSBO seller would
perceive a “guarantee that, in addition to listing their property for free on
Zillow’s website, no real estate agent could represent a potential buyer or be
involved in any part of the potential transaction,” and therefore chose Zillow
over Picket Fence. [You can tell how this is going to go by the description of
the claim.]

The basic problem: “Zillow’s advertisement would not mislead
a consumer who was interpreting the message reasonably. Zillow simply
advertises that FSBO sellers can post for free, and that is true.” It made no
additional representations that Premier Agents would not be involved in FSBO
transactions, nor that FSBO buyers would not incur transaction costs. The
complaint even alleged that sometimes buyers of FSBO properties have agents,
which made Picket Fence’s interpretation unreasonable.

A predatory pricing claim also failed, as did an argument that
FSBO ads must contain no agent/intermediary involvement. This was an implicit
falsity argument, and the allegations in support of the consumer deception
element were:

Picket Fence had former Zillow For
Sale By Owner customers complain about the deception on Zillow and specifically
said had they known the truth about how Zillow operated its website and their
For-Sale-By-Owner advertisements, they would have chosen to advertise with
Picket Fence. Picket Fence is aware that For-Sale-By-Owners would ch[o]ose a
free advertisement … since they assumed a free advertisement [ ] would be a
cheaper alternative.

But these were conclusory allegations without supporting
detail, and anyway there wasn’t even implied falsity, since Zillow wasn’t
alleged to have required the use of an agent by FSBO sellers.

from Blogger http://tushnet.blogspot.com/2023/08/allowing-free-for-sale-by-owner-ads.html

Posted in Uncategorized | Tagged | Leave a comment