CopyrightX: UGC panel

Panel One: User-generated Content, Digital Labor, and Collaborative Authorship

Moderator: Bethany Rabe

Rebecca Tushnet: Fanworks, Fair Use, and Self-Actualization Through Transformative Expression

Title assigned a few months ago is a little misleading because I actually wanted to take the opportunity to talk about the Copyright Office’s recent report on moral rights, though I’m happy to talk about anything fanwork related during the panel discussion and Q&A.  I was one of the founders of the nonprofit Organization for Transformative Works.  OTW runs the Archive of Our Own, which hosts noncommercial transformative works by fans of existing works, of people, or really of anything about which one might create new art. From our recent fact sheet I can tell you that we have over 1.2 million registered users and 4.7 million fanworks, as well as tens of millions of visits per week. “Our open-source code was built from the ground up by fans for fans…. Our users are fans of every sort—teenagers and grandparents, first-time writers and professional authors—from all around the world, creating fanworks” for nearly 32,000 different fandoms.  From the beginning, it was a priority to be ad-free and noncommercial, not under the control of advertisers.  Instead, the priority was authorial control: AO3 was designed to give creators the ability to post and edit their works flexibly, to orphan their own works to preserve access to them without connecting them to the author, and to exclude search engines or non-Archive users from reading.

One of the most distinctive features of the Archive is the curated folksonomy used to tag individual works: “Users are able to tag in whatever format is most useful or natural to them, and our team of over 350 tag wranglers link these tags together into easily searchable concepts.”  If you misspell Harry Potter’s name—or put it in kanji—we’ve got you covered.  Creators can use tags to allow other users to find exactly what they want to see—or to exclude exactly what they don’t. “[U]sers can easily create specific searches, including all Sherlock Holmes works posted in 2018 that are exactly 221 words long and Lord of The Rings/Game of Thrones crossovers that don't include either Frodo Baggins or Arya Stark.” “Related concepts are also linked: Space Opera and Space Battles are both found under the metatag Outer Space, along with other related terms like Astronauts, Spaceships, and even Space Whales.”

Fanworks provide incredible benefits in literacy, sexuality education, language learning, community building, and other good things that come from making something you love and sharing it with other people who might just love it too.

On the content v. tech distinction that Maria Strong made: we consider ourselves content folks who use tech.  But importantly, we rely on fair use: our basic purpose is to host transformative noncommercial works. Gives us a particular perspective on various copyright controversies.  For example, recent Copyright Office report on moral rights, which you’ve just heard about.  The Office did a lot of good work pulling together the legal background both internationally and in the US, covering many issues that go beyond copyright into other regimes that affect authors.  A moral right of integrity that would protect authors against uses of their works they found offensive would be incompatible with US fair use and the First Amendment and the Office didn’t suggest any basic changes to that system, to the Office’s great credit.

It is definitely true that norms of fairness matter a lot in authorship, including attribution in many circumstances—but not necessarily in the rigid categories formal law might use.  For example, fans regularly describe authors of books, but generally identify performers of songs rather than songwriters. More generally, attribution practices in and out of fandom are highly context specific—what works for one group might not work for another.  Hollywood screenwriters have an arbitration system for assigning credit that is often needed because fights regularly break out.  Similarly, the Office pointed out that, “in legal writing, attribution norms for academic articles are quite rigid, whereas practicing lawyers routinely copy without attribution ‘the form and language of legal instruments.’” Attribution turns out to be the kind of inquiry that isn’t well suited to the American legal system, because modern copyright statutes and regulations tend to make big, specific lists of required information.  The Office wisely recognized that antiplagiarism norms work about as well as anything can work, and that the context of attribution would be essentially impossible to write into law.

Nonetheless, and without showing that there was a noticeable unmet need for new protections, the Copyright Office has just suggested that the federal Lanham Act can & should be used against unattributed copying of digital works, in defiance of the Supreme Court’s decision in Dastar and its own stance on digital first sale.  [I should note I think the Office is completely right in saying (1) §43(a)(1)(B) claims should survive Dastar where false attributions are material to consumers and (2) Dastar applies to in-copyright works as well.  It’s just disingenuous at best in endorsing a distinction between nondigital and digital copies that does not make sense of the Supreme Court’s interpretation of “origin,” especially in light of the Office’s (correct) position in discussions of first sale that a digital copy stored on a computer is a reproduction.  The “origin” of a digital copy is the host computer, and maybe stretching it could be the sending computer, but there is a physical instantiation and that instantiation is what the Lanham Act covers in “origin” under Dastar.  Specifically, the origin of the creative work contained in that copy is not the same thing as the origin of the physical copy, no matter whether the copy is on a general-purpose computer, a videotape, or a DVD, and thus failure to attribute the origin of the creative content is not actionable under §43(a)(1)(A) per Dastar.]

It also stated that Congress might want to consider amending the Lanham Act to create new causes of action to cover false representations regarding authorship of expressive works, even without a showing of harm to consumers. The Office suggested amending the Copyright Act to provide damages when a defendant knowingly removed or altered copyright management information (“CMI”) with the intent to conceal an author’s attribution information even if there was no intent to conceal infringement, and possibly even adding a federal right of publicity.  In addition, the Office suggested some changes to the more limited Visual Artists Rights Act. 

Aside from the suggestions about VARA, the rest of the Office’s ideas are a classic case of proposing to use a missile against a gnat—likely to the detriment of small and noncommercial creators, who are the ones who can neither afford to bring claims nor fight claims brought against them.  And the specific suggestions lack exactly the kind of context sensitivity the Office acknowledged was needed—for example, the proposed CMI provision on its face makes it illegal for an app to remove identifying information of people submitting anonymous videos of police brutality—and if you think that enterprising prosecutors wouldn’t use this law against the ACLU, you haven’t been paying attention.  [The CO says that the intent requirement would mean that automated removal wouldn’t be actionable unless it was designed to remove authorship information … but anonymous reporting apps are designed to do that. And for what it’s worth, the complaints the CO reported receiving about removing authorship information were mostly about automated processes—the proposed remedy doesn’t even match the thing that’s being complained about.]

It’s also worth pointing out the solution that the Copyright Office didn’t suggest for congressional intervention: although its discussion of authors’ rights mentioned attribution and integrity rights secured by labor law and collective bargaining, both in the US and elsewhere, the Office didn’t suggest that strengthening authors’ positions could be achieved by strengthening the position of workers, even though the most vulnerable authors are labor and not management.   The problems the Office identified in securing attribution for certain groups, like freelancers, are labor problems—authors after all make copyrighted works, and we should not allow the concept of the “work” to disappear as if authors were mystical beings with no connection to other forms of labor, other forms of production, that do things in the world.  Our desire to highlight the process of creation was in fact one of the reasons that we called our nonprofit the Organization for Transformative Works: whether made for profit or not, creative works are the product of labor.  Remembering the inherent dignity of that labor, and its value whether or not there is a market for the resulting creativity, is a goal that the OTW shares with the Copyright Office.

Stacey Lantagne: Internet Memes, Fanworks, and Copyright’s Authorship Challenges

Memes regularly involve © works; one opinion expressed: the public “owns” the meme because the public authors the meme even if not the underlying photo.  Fox defended an infringement case juxtaposing 9/11 and WWII images—court didn’t like the fair use defense b/c didn’t see the transformativity, but it turns out that the range of “memes” is very broad.  Knowyourmeme has no definition.  When Trump (or a Trump fan) adds CNN’s logo to a clip of wrestling so that it looks like Trump is beating up “CNN,” that’s what’s adding the meaning especially after it’s been shared by Trump himself. Similarly, Pepe the Frog became a white nationalist symbol and created contests over ownership w/initial artist and those who altered it. Gab now uses a frog logo and claims that frogs have been symbols of free speech for a long time (narrator: they have not), and it’s a reference but not all that similar to Pepe.  Then the distracted boyfriend meme, which comes from a stock photo with a © owner, who is fine with most uses but not w/objectionable ones (which may be censorial).  As the distracted boyfriend meme develops, responses to it/reworkings of it are based on transforming the meme rather than on the photo alone, which has become subsumed into the meme and the idea of the meme. 

The internet thus exposes that the mastermind narrative of single creation is wrong. We knew that but dealt with it mostly by contract, and then we have this control test for noncontractual situations but it still wrongly assumes that there is a “boss” in control of everything. When you don’t have contracts, when you don’t have lawyers, we don’t know what we’re doing in terms of ownership. Garcia v. Google: opinions don’t make sense internally and are conflicting all over the place.

Capitalism makes everything worse/more exploitatitve: Swedish company used distracted boyfriend meme for job searching and ad regulator called it sexist and said it couldn’t be used in ads. There is a pending case against a MAGA poster by Pepe’s original creator Matt Fury; there’s a fair use argument but also the defendant argues that Fury abandoned the © (which is not a thing).  There’s a lot of unsettled territory here.

Hilary Richardson: Google Open Source and Collaborative Authorship

Writing a casebook on open source legal issues: opensource.google.com/docs/casebook: you can send us edits or additions in a GitHub pull request.  Current joint authorship standards give equal rights to all authors.  Some projects list all “authors” giving them shared control. Some may have 100% consensus, but for a growing project that can be unsustainable, and some institutions may not want the potential liability/disputes that comes from open source authors’ potential for contestation.  Current doctrine (Al-Muhammad etc.) looks for a mastermind—important contributions can still avoid being authorial. Court was concerned about discouraging future editing/feedback if joint authorship was a possible consequence. Creative control/direction mattered, versus value of contributions and impact on final version. Spike Lee had the control (although it was a work for hire) and the court found that determinative.  Other jurisdictions also use editorial control over final piece as part of the test, but with more focus on intent to be co-author.  Control can be evidence of intent, but intent could also be evidenced by, e.g., registering the © in a person’s name.

Clarity can come from legal agreements: two popular mechanisms, developers’ certificate of origin (Linux) which doesn’t say anything about retaining © but does grant an explicit license.  Is everyone a joint author?  Contributor license agreement: says that © stays with contributor but they grant a license.  That makes it clear that people aren’t considered joint authors.  More light-weight approaches: root license of a project is used to cover contributions. Can be used to imply effects on individual contributions; sometimes licenses say “if you contribute, you do it on these terms.”  The Stuff case about contributing special effects to a larger film: helps explain the effects of an implied license to use a contribution.

Katrina Geddes: Should User-Generated Content Be Compensated?

UGC: just about everything you find on the internet.  Childish Gambino’s This is America was a canrivalesque commentary on police brutality, fetishization of black performers, modern-day minstrelsy. People responded w/, e.g., This is Nigeria: commentary on police brutality, corruption, poverty, Boko Haram, and other Nigerian issues. Similar music and choreography but different lyrics (both set in warehouse).  Why do we value things like fan fiction?  People get to see themselves where they don’t get to see themselves in mainstream works: LGBTQ people, people of color. Challenge racial and social assumptions of mainstream works; challenge the monopoly on cultural production of large producers. Can also maintain the cultural relevance of works, such as maintaining interest in the Star Trek universe.  Satisfies demands for variation and personalization, rather than works manufactured for a huge audience. Democratizes creativity—authorship is diverse and diffuse.  Meme authorship can often be indeterminate. 

Content ID: used algorithmically mostly for commercial content.  400 hours to YouTube uploaded per minute. Fanmade videos for Harlem Shake were matched to the song, allowing the song © owner to generate millions of revenue with no allocation between the music and the video created by fan labor. Under 1% of Content ID matches are disputed.  Fair use can’t reliably be done by algorithms; Content ID is black box governance, unaccountable—even when the uses removed or monetized may be perfectly lawful.  Expands scope of © by making it difficult for users to rely on exceptions like fair use.

Locke and labor theory: just reward of labor should extend to fan creations that are monetized.  Eric Posner etc. have argued that not just UGC but our labor online, clicking and producing valuable data, should be understood as compensable.

Canadian UGC provision: protects UGC made by an individual, though there’s usually more than one author.  Also protects UGC only when © subsists, which bakes in originality.  Requires noncommerciality—no pro-rated advertising revenue.  Also bars anything with a vague, broad “adverse effect, financial or otherwise” on the © owner—that could include offending the original © owner.

Should users be compensated?  The labor of users should be recognized and respected, and possibly compensated. But commercialization is part of the fair use inquiry; it may weigh against fair use.  Commercialization may result in reduction in expressive diversity: anticipation of paid markets may affect content and willingness to participate/be spontaneous in creativity.  Distinguishing when compensation would be paid would require resources, negotiation in allocation between different owners. Right now it’s 100% to the original claimant, but other points on the spectrum are possible. Don’t just assume that everything with a match should be 100% monetized by the © owner.  We could ask users to categorize their works to claim a share; this would allow more information and an algorithm could at least look for percentages/set things up for human review.

Q: fair use is important; from litigator’s perspective, one downside is expense to determine.  Oh, the Places You’ll Boldly Go! was found to be fair use on summary judgment, but not at the MTD stage which was 1½ year earlier.  Is there a way to get to fair use faster?  Market harm may require discovery—that’s the most expensive part of most lawsuits.

RT: sadly, not a lot.  There are uses that are fair as a matter of law, including parodies.  Big data, purpose-transformation uses are for better or worse somewhat easier to identify as fair use as a matter of law.  Sometimes you are going to have to evaluate the transformativeness of an individual work.  [We can have rules that help, like the Second Circuit’s rule that you can’t avoid a fair use finding by offering  a license for something that is in fact transformative—that helps draw the sting of some market evidence.] Small claims: 40% of Americans can’t come up with $400 for an emergency.  A “small claims” court that allows $30,000 in statutory damages is not a small claims court. It’s just another way to harass individuals.  [I forgot to add: best practices in fair use for different fields—also very helpful.]

Lantagne: education, nonprofits like OTW/EFF.

Geddes: Canadian UGC exemption needs some time to develop.  Education is also important.  People want rules like “30 seconds is ok.”

Richardson: people do want rules: are 3 lines of code ok?  Depends on what the 3 lines are!

Kimberly Isbell: proposed 1202A.  The reason we made it a new provision is b/c it’s narrower and specifically exempts criminal enforcement.  The only person who could bring a claim would be the uploader. The ACLU has terms of use: if you provide content it will be anonymous.

RT: Disagree.  The statutory language is “any person injured by a violation,” and the police will say they’re injured by the inability to investigate the circumstances under which the video was taken further.  The injured person doesn’t have to be the author (or heirs) under this proposal.  There is a long tradition of (mis)using civil awards to silence speakers, e.g. NYT v. Sullivan.  The proposal does not exempt information submitted by someone who wants to be anonymous or who doesn’t know or care about stripping of metadata [nor does it limit enforcement to the author, which would also be a potential fix, though I still don’t see the evidence that an additional cause of action is needed].

Q: software should just be patentable, not copyrightable.

Richardson: the law says copyrightable and she’s ok with that.

Q: memes & defamatory content. There are a lot of really awful Helen Keller memes out there.  What should happen when there’s a historical figure & derogatory content.

Lantagne: a lot of times people bring © claims that should be more personal tort based.  She hasn’t come across non-© meme cases, but she wasn’t specifically looking for them.  Unless publicity rights outlast death, there are few remedies for the dead.

Richardson: revenge porn is an example of where © was deployed even though it wasn’t intended for those situations; you shouldn’t need to register an explicit photo to deal with the problem of revenge porn, and new laws and platform policies are recognizing that.