Tag: electronic frontier foundation

It Looks Like the EFF is Pro FAKES

FAKES

When it comes to cyber policy and anything like intellectual property, the Electronic Frontier Foundation’s critiques are so predictable, they might as well use ChatGPT to write their blog. For instance, in opposing the NO FAKES Act, an April post by Corynne McSherry selects items from that same menu of responses EFF has used to oppose any form of online copyright enforcement. In this instance, she orders up the following:  pretend to want a “better” bill; cite scary hypotheticals; pretend to care about creators; and, of course, insist that the speech right is in jeopardy.

For review, the NO FAKES Act would establish a new property right in every individual’s likeness, including one’s voice. As opined on this blog, its mechanisms comprise a thoughtful response to a novel challenge—namely the ability of just about any party to use generative artificial intelligence (GAI) to replicate the likeness of any person. The hazards of replication are obvious to the common-sense observer—from intensifying disinformation to commercial uses without permission to sexual predation, scams, and harassment. But as usual, the EFF advocates the interests of the tech industry by framing its critiques in a rhetoric that sounds pro-individual or even (ha!) pro-artist.

McSherry’s broadside at NO FAKES employs the tactic of alluding to hypothetical negative consequences, which Congress has (of course) failed to consider. Thus, EFF insists, as it did with bills like the CASE Act, that NO FAKES, as written, should be balled up, and that Congress should start over from scratch. But those of us familiar with the organization recognize that this pretense is there to mask the EFF’s view that the whole idea of a likeness right should be scuttled. If past is prologue, the EFF will never endorse any version of a law to remedy unlicensed AI likeness replication and, possibly, never engage as a good-faith negotiator on the subject.

Predictably, McSherry’s post elides important details about NO FAKES. I won’t unpack them all, but in one example, she writes, “The right applies to the person themselves; anyone who has a license to use their image, voice, or likeness; and their heirs for 70 years after the person dies.” Of course, she doesn’t mention that although the 70-year term is the maximum, the likeness right would have to be renewed post-mortem and, similar to trademarks, renewal is conditioned on showing that the likeness is still in “authorized public use.”

But it was another paragraph that struck me as vintage EFF—an implication that existing right of publicity (ROP) laws in the states are already harmful to speech and that NO FAKES can only make matters worse. McSherry writes:

… it’s become a money-making machine that can be used to shut down all kinds of activities and expressive speech. Public figures have brought cases targeting songs, magazine features, and even computer games. As a result, the right of publicity reaches far beyond the realm of misleading advertisements and courts have struggled to develop appropriate limits.

NO FAKES leaves all of that in place and adds a new national layer on top, one that lasts for decades after the person replicated has died.

But following the links in that first paragraph, one finds a couple of unmeritorious claims of ROP along with a couple of EFF’s opinions about how ROP law should be applied in context to the speech right. In the first instance, weak cases that do not prevail only disprove the allegation that a law has “reached beyond” its intent. And in the second instance, while the EFF is entitled to its opinion, its interpretation of the speech right is so expansive that it is unremarkable when the courts so often disagree with their positions.

In fact, the EFF’s overbroad concept of the speech right is one reason I say it is being disingenuous in asking for a revised likeness bill. NO FAKES arguably provides better guidance on the use of AI replicas for protected speech than ROP case law, but although McSherry acknowledges these provisions, she states, “…interpreting and applying those exceptions is even more likely to make a lot of lawyers rich.” That’s code for “let’s not have anything like this law” because of course all laws need to be interpreted and, yeah, lawyers are usually involved.

Another hypothetical includes the familiar, and laughable, implication that the EFF cares about creators and performers…

People who don’t have much bargaining power may agree to broad licenses, not realizing the long-term risks. For example, as Jennifer Rothman has noted, NO FAKES could actually allow a music publisher who had licensed a performers “replica right” to sue that performer for using her own image.

While it is true that an individual could over-license the use of her likeness to another party, this is no different than licensing traditional forms of intellectual property. That an owner might give away too much is a consideration of the owner’s savvy and legal representation, but not a rationale to oppose the right being established in the first place. This complaint is a rehash of the fallacy that copyright rights are bad because some parties have cajoled artists into signing over more than they should. As applied to NO FAKES, I suspect people will favor the right to control their own likenesses and then worry about licensing, if that becomes an issue.

Finally, McSherry’s post repeats the same old prediction that NO FAKES will lead to platforms removing some undefinable, yet unacceptable, volume of protected speech. It’s almost surprising that EFF remains committed to this message when social media is clearly overflowing with so much protected hogwash—and when the major platforms are increasingly taking down innocuous posts without any rationale or transparency. A casual review of the current state of “information” on social platforms can only support the rational prediction that AI generated likenesses will exacerbate the problem. At the same time, EFF’s claim to defend “new creavity” is overstated when even protected uses of AI likenesses are often little more than brief diversions of limited cultural, and no informational, value.

For every potentially legal use of AI likeness, there are dozens of ways for scammers, foreign adversaries, predators, and unscrupulous business operators to use the technology to cause serious harm. But, true to form, the EFF asks that we ignore evidence of the damage being done and imagine instead that any remedy must be worse than the disease. Just off the cuff, they’ve used similar tactics to be wrong about CASE Act, Section 1201, Section 230, site-blocking, and Controlled Digital Lending. So, it is hardly a bold speculation to say that they’re wrong about NO FAKES.

Image source by: maxxyustas

The EFF Campaign Against DMCA Section 1201 Perishes in the DC Circuit

section 1201

The First Amendment protects the right to read books but not the right to break into a bookstore for the purpose of reading—not even if the goal is to quote a passage from a book in a manner that would be fair use under copyright law. The hypothetical, lawful use of the book’s contents to produce protected expression does not make the law prohibiting trespassing into the store a violation of the speech right. Most reasonable people can understand this distinction, but for about 18 years, the Electronic Frontier Foundation (EFF) has tried to prove that common sense is wrong.

Ever since literary works, sound recordings, audiovisual works, etc. went digital, the concept of “digital locks” used to protect lawful access to these materials has vexed the EFF, which launched a campaign and lawsuit in 2016 to prove that the law against breaking said locks is unconstitutional. Filing suit on behalf of researcher Matthew Green and product developer Andrew Huang, the EFF has argued that Section 1201 of the Digital Millennium Copyright Act (DMCA) violates the speech right because circumvention of technical protection measures (TPM) may sometimes be done to achieve forms of protected expression that would be defensible under the fair use exception.

TPMs generally consist of code used to enforce lawful access to digitally distributed works like eBooks or streaming services, and §1201 prohibits circumventing TPM and/or trafficking in devices primarily designed for circumvention. By law, the Librarian of Congress (really the Copyright Office) conducts a triennial rulemaking proceeding to consider applications for, and grant exemptions to, §1201 for purposes such as research and certain educational or journalistic uses of the encrypted works. You can read posts here and here for background on the EFF’s case, but the bottom line is that the appellate court last week soundly rejected the claim that 1201’s “legitimate sweep” functions as a “speech licensing” law.

Among the court’s determinations, it held that the government’s interest in preventing “digital trespass” properly restricts a wide range of conduct that has no expressive purpose; the First Amendment does not guarantee unfettered access to expressive works; the plaintiffs in name make no showing that their protected expression is being chilled; and the various hypothetical examples presented by the EFF are answered by legal forms of access to works that do not require circumvention of TPM. More than a few of the court’s responses demonstrate why the EFF has tried inaptly to portray an anti-trespass law as a speech law. For example…

A trespass law undoubtedly affects some expressive conduct, as when political protestors trespass to stage a demonstration where it might have maximal impact. Similarly, the DMCA’s anticircumvention provision might preclude a student from circumventing technological measures to cut a high-quality clip of a copyrighted feature film to use in his class presentation. But trespassing is not “necessarily associated with speech,” because laws prohibiting trespass also “apply to strollers, loiterers, drug dealers, roller skaters, bird watchers, soccer players, and others not engaged in constitutionally protected conduct.”

As the court explains, §1201 likewise applies to a range of parties with an interest in circumvention for both lawful and unlawful purposes, but expression is not the basis on which the law operates. Going back to the bookstore, it is simply illegal to break in at all, regardless of whether the intent is to read, cite a book, or ransack the place. The fact that the vandal will face charges not attributable to the reader has no bearing on the trespass violation they both committed.

I also want to highlight the court’s response to the allegation that the §1201 rulemaking process is itself an unconstitutional prior restraint on speech as indicative of EFF’s chronic misstatements about fair use. The court writes, “An irony of appellants’ challenge to the DMCA is that the triennial rulemaking exemption scheme—which identifies in advance and immunizes categories of likely fair uses—may be less chilling of the fair uses to which it applies than the after-the-fact operation of the fair use defense itself.”

In other words, arguing a fair use defense requires litigation and uncertainty in contrast to a rule by the Librarian that a given use has been granted an exemption. The Library has granted a broad range of exemptions to §1201, and as this opinion notes, an exemption granted to a single petitioner (e.g., a documentary filmmaker or teacher) applies to all parties in that class with the same interest in circumventing TPM.

Finally, the court concludes that the rulemaking proceeding is not above judicial review—that a petitioner who believes the Library has made a content- or viewpoint-based decision may still bring a First Amendment complaint to the courts on that basis, but this does not alter the finding that the law itself withstands constitutional scrutiny. Never say never, I guess, but I predict this alleged controversy is now a settled matter—that EFF has wrung all the value it can from this campaign and will need to find a new anti-copyright windmill on which to break its lances.

DMCA Section 1201 Still Constitutional Says Circuit Court

The Electronic Frontier Foundation (EFF) was dealt a significant (possibly fatal) blow in its longstanding endeavor to have the courts abolish the entirety of DMCA Section 1201 as an unconstitutional violation of the First Amendment. The case Matthew D. Green, et al. v. United States Department of Justice was filed in July of 2016, and on December 6, the DC Circuit Court of Appeals affirmed the lower court’s denial of a preliminary injunction against enforcement of 1201 on behalf of plaintiffs Matthew Green and Andrew Huang.

Section 1201 is the provision within the Copyright Act that enables rightsholders to use computer code for the purpose of digital rights management (DRM) with two primary purposes: 1) to control access to works; and 2) to prohibit unauthorized copying of works. These purposes naturally work in tandem. For instance, Netflix requires a subscription for access and allows an account holder to download a file for offline viewing, but different code prevents copying that file for transmission to some other party. The 1201 statute prohibits unauthorized circumvention of DRM and/or trafficking in technology that enables unauthorized circumvention or copying, either of which may result in civil or criminal action.

Simply put, without DRM, the digital, on-demand market for most media would not exist, and the central purpose of 1201 to foster and sustain this market is a settled matter—both in Congress and the courts. Nevertheless, the EFF has vowed to see 1201 abolished in our lifetimes. In 2018, in a blatant appeal for donations, Cory Doctorow wrote a blog for EFF touting its “Apollo 1201 Project [which] aims to kill all the DRM in the world inside of a decade.” No mention of the fact that, at that point, the EFF’s original endeavor to “kill” DRM with the Constitution was already about twenty years old. And as of last week, it is fair to say that if DRM is to die circa 2026-28, it will not be at the clumsy hand of the EFF and its flawed arguments on behalf of inapt plaintiffs.

In the case of Mr. Green, he was seeking a preliminary injunction as a prophylactic shield in advance of publishing a book which explains (in both English and computer code) how circumvention is achieved. But the appellate court states:

Because Green intends to provide code able to circumvent technological protection measures, he believes the book would likely violate the antitrafficking provision. At oral argument, however, government counsel made quite clear that in its view, Green’s proposed course of conduct would not run afoul of the DMCA…. The government’s concession ends any “credible threat of prosecution” against Green, leaving him without standing to obtain a preliminary injunction.

As my friend Devlin Hartline mentioned during a call to discuss this case, “It was surprising to see the government concede that ground here, but even if it had not, Green’s argument for a preliminary injunction would have lost on the merits.” Writing a book about circumvention is protected speech that does not implicate 1201’s anti-trafficking provision, and the extent to which said book might be construed as a How-To manual for circumvention, that would be a fact-intensive inquiry which would still be tough to argue is a violation of the statute. This may explain why the government shrugged at it in this case.

As for Andrew Huang, he sought an injunction for intended activity that goes to the heart of 1201. Hoping to sell a DRM circumvention device called NeTVCR, along with published code so the device can be updated, Huang (the EFF) argued that computer code, even for circumvention, is protected speech. To this, the circuit court essentially said “no kidding” and notes that the government never challenged the premise that code itself is speech under the First Amendment. But as the opinion states, “We turn then to whether the DMCA ‘target[s] speech based on its communicative content’—that is, if it ‘applies to particular speech because of the topic discussed or the idea or message expressed.’”

Separating Expression from Function

“The Supreme Court’s recent free speech case, City of Austin v. Reagan National Advertising of Austin, LLC, 142 S. Ct. 1464 (2022), is virtually dispositive,” the DC Circuit opinion states. At issue in Austin, and the relevant caselaw, is the distinction between state sign regulations that are content-neutral (and likely not in conflict with the First Amendment) versus regulations that may be content-focused (and likely to conflict with the First Amendment). The City of Austin’s rules pertain to off-premise signs, which may advertise, promote, or inform readers about something unrelated to the location of the sign; and on-premise signs, which may advertise, promote, or inform readers about something specific to the location of the sign.

Without getting into the weeds on various state sign regulations, suffice to say, the SCOTUS majority in Austin held that the city’s regulations at issue are content neutral, rejecting the petitioner’s argument (and the dissent) that because enforcement necessitated reading the signs, the regulations are fundamentally content based. The majority held, “Underlying these cases and others is a rejection of the view that any examination of speech or expression inherently triggers heightened First Amendment concern. Rather, it is regulations that discriminate based on ‘the topic discussed or the idea or message expressed’ that are content based.” Making a comparison to Austin, court in Green states:

The same logic applies here. Although the DMCA requires reading computer code to determine what digital act the code carries out, it is nonetheless content neutral because, in the words of City of Austin, it cares about the expressive message in the code “only to the extent that it informs” the code’s function.

The court further states that the DMCA question presented is easier than Austin because that case concerned the regulation of speech as speech, while Section 1201 is only concerned with the functional aspect(s) of computer code. As any copyright nerd is aware, the copyrightability of code is acutely concerned with reconciling the expressive (protectable) from the functional (non-protectable) aspects of the work. And here, as elsewhere, that which is copyrightable (expression) happens to be the part of computer code that would be protected speech. The EFF wants to call locks and keys forms of expression just because they are made of code instead of steel. But, of course, the locks are made of code — because so are the works they protect against theft.

Devlin Hartline, writing about this case for Copyright Alliance in September of 2021, discussed EFF’s alternate (reverse engineered?) route to a First Amendment challenge to 1201—namely that DRM is a prior restraint on potentially non-infringing conduct. The organization has argued many times, for instance, that if the intended use of the copyrighted work may be a fair use, then 1201 prohibits this otherwise legal conduct, which is itself protected by the First Amendment. But, as Hartline plainly observes, “The First Amendment does not allow you to pick the lock on the front door of a library in the middle of the night in order to break in and make fair use of the copyrighted books in the stacks.”

One does not need to be a constitutional scholar or a computer programmer to grasp the common sense understanding that the provisions of 1201 address solely functional aspects of code. In plain terms, 1201 is silent as to what code may be used for DRM, and it is likewise silent as to what code may be written that could pick the proverbial lock. It simply states that the act of picking the lock or trafficking in lock picks is against the law for the same reasons that bootlegging was illegal in the days before digital.

To work at the EFF, it seems that one must swear a blood oath to the core belief that copyright is inherently in conflict with the speech right. And orthodoxy mixed with righteousness (and more than a little self-interest) will invariably produce convoluted logic. Since 2001, the organization and its likeminded amici have failed to find any purchase in the courts with a constitutional challenge to DMCA Section 1201. As opined in an older post about this case, if EFF et al. really cared about specific burdens on certain parties, it would put more energy into advocating relevant permanent exemptions via the Copyright Office rulemaking process. Why it persists in this boondoggle of endless, ill-fated litigation can best be explained, I think, by the “Donate Now” button.