Thursday, August 03, 2023

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

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.

No comments: