Category: Intellectual Property

DMCA Review Should Be About Copyright, No?

On September 30, the House Judiciary Committee held a hearing to discuss the Copyright Office report, published in May, commenting on the efficacy of Section 512 of the Digital Millennium Copyright Act (DMCA). Section 512 provides conditional immunity to online service providers for copyright infringements conducted by users of their services. (For a basic summary of conditions, see page here.)

Reiterating the position that the USCO report fails to consider the interests of the general public in its analysis, Meredith Filak Rose, senior policy counsel at Public Knowledge, urged the committee to proceed with cautious awareness that in the years since 1998, the public has become profoundly dependent upon the internet for a broad range of ordinary and essential needs.

With due respect to Rose personally, and with deference to the many devil’s details implicated by her testimony, I shall, once again, take issue with the over broad context in which digital rights groups like Public Knowledge try to frame discussion about the DMCA. For instance, at the start of her testimony Rose states that, “229 million Americans use the internet each day. That’s 229 million American adults using the internet to work, worship, connect with family and friends, receive healthcare, consume and discuss the news, and organize political action each and every day.”

Aside from the fact that the mosaic of internet uses needs to be more diverse in order to present a clear picture (let’s not forget the mindless scrolling, the clickbait, the misinformation, or the porn), the salient point is that most ordinary internet use does not require the appropriation of copyrighted works. So, framing a conversation about a section of the copyright law by alluding to the scope of everyday internet traffic is both distracting and entirely beside the point. If Congress were discussing CAFE standards, and an oil industry representative testified that 229 million American adults drive to work, church, and the grocery store every day, this would be a meaningless prelude to an argument against mandates for more fuel-efficient cars.

The Fight Over Account Termination

So, let’s stipulate the obvious:  We all use the internet for myriad practical purposes all day long. And if anyone can show me the intersection between copyright infringement and a telemedicine appointment, I’ll take a look. But what Rose is really teeing up is advocacy for the status quo of DMCA §512(i) and the barely implemented requirement that ISPs eventually cancel the accounts of repeat copyright infringers. We cannot reconcile, Public Knowledge argues, a family’s fundamental need for broadband with the possibility that a teenager in the house might repeatedly infringe copyright, and the service provider will be required to terminate access for the entire household.

But the reality is not quite so binary or draconian, even if the statute has proven unclear to the point of futility. Congress’s decision in 1998 not to define “repeat infringer,” or to codify universally applicable guidance for termination policies, left the ISPs (access providers) and the edge providers (web platforms) free to maintain the practice of termination avoidance for repeat infringement by users. The concern of digital rights groups, therefore, is that somehow the service providers will have to comply with a 22-year-old condition they’ve largely evaded.

In the costly litigation COX v. BMG and that provider’s risible 14-strike policy, COX’s users received multiple warnings before not actually losing their accounts. And although copyright owners would certainly like to see more meaningful implementations of 512(i), they neither propose nor endorse a scenario in which a family wakes to find its broadband inexplicably terminated for repeat infringements of which the account holder was somehow unaware. This is not the way account termination happens now or has ever been envisioned to happen.

At the same time, although this is not the post for offering specific legislative recommendations, one policy that would alleviate some of the tension in 512(i) is site blocking, which has proven effective in foreign jurisdictions. If groups like Public Knowledge, EFF, et al were not so adamantly opposed to blocking enterprise-scale, foreign-based piracy sites, a compromise might be more easily found that would mitigate many of the concerns these groups identify with regard to account termination scenarios.

“Red Flag” Knowledge at the Heart of the Matter

This focus on the internet writ large reinforces the major internet companies’ efforts to conflate their commercial interest with the public interest. What many call the “free flow of information,” allegedly for our benefit, often has nothing to do with information. What this erudite sounding expression really means is that because the social sites are engineered to exploit vulnerabilities in human psychology in order to keep users addicted and active, the platform owners like to avoid legal obstacles like copyright, privacy, or anti-trust matters that may create friction between user and interface.

Consequently, today’s major platforms—all founded years after the DMCA was first hammered out between big telco and big media—read certain ambiguities in the statutes to mean that they are free to profit from chronic infringement by users, while doing the bare minimum to comply with the notice-and-takedown provision. Specifically, as discussed in my post about the first Senate-led review of DMCA, rightsholders hope that Congress will more clearly define §512(c), which states that providers will not be liable for infringement if …

(1) its operators do not have actual knowledge of infringement; (2) its operators are not aware of facts or circumstances from which infringing activity is apparent; and (3) upon obtaining knowledge of infringement, expeditiously removes the relevant material.  

Commonly referred to as the “red flag” knowledge section of the statute, a major point of contention for rightsholders, both in and out of court, is the extent to which service providers allege that they lack any knowledge of infringement sufficient to meet the liability standard. Even in a relatively recent case where plaintiffs presented emails that revealed site operators made affirmative decisions to leave material online they believed to be infringing, courts have misread §512(c) to mean that these operators would need legal and industry expertise to meet the “red flag” bar. This is inconsistent with the reasonable, ordinary person context in which this part of the statute was written, hence the hope by rightsholders that Congress will consider clarifying the language.

Because §512(c) is at the heart of the good-faith/shared responsibility intent of the DMCA, I have to say that I did a little spit-take when Rep. Lofgren raised the “red flag” subject and asked her first question of Meredith Rose, who replied that she is “not terribly familiar” with that part of the statute. This is not intended as a personal gotcha, but it is a rather serious matter when an organization purporting to represent the interests of “everyone who uses the internet” is unprepared to discuss one of the most problematic sections of the DMCA. In fact, the much broader question of what platform operators can know about the material on their servers, and what they should do about some of it, is the vexing challenge of the moment with regard to the effect social media are having on society. The knowledge question goes way beyond copyright.

The Dogeared Speech Argument

Historically, the internet industry’s shell game on the subject of what can and cannot be known is consistent with the kind of site management that has now proven to be the major catalyst in the dissolution of democracies worldwide. The same companies whose algorithms are allegedly so sophisticated that they can predict our choices before we make them, paradoxically claim an inability to parse data that ordinary, non-prescient humans can interpret. The manner in which the industry has exploited vagueness in the knowledge standard in the DMCA runs parallel to its history of shrugging “neutrality” when it comes to the moderation of harmful material like organized hate speech, conspiracy groups, and dangerous misinformation—a “neautrality” no longer acceptable to much of the public.

I cannot fathom how any reasonable person looks across the landscape at the ragged state of American democracy and, with a straight face, continues to exalt Web 2.0’s grand experiment in free speech as though it were not an appalling failure. The evidence is now clear, including testimony from a steady stream of defectors from the social media companies, that Facebook, Google, Twitter, Reddit, et al purposely designed their platforms to be digital crack. And it is no surprise that divisive politics and conspiracy garbage are potent ingredients in the drug cocktail that captures and retains the attention of millions.

Referring back to the 229 million users, it isn’t connecting to family or online banking or worshipping that is systematically destroying the American Republic; it’s the speech-a-palooza that organizations like Public Knowledge earnestly champion that has sown a motley patchwork of customized realities to the extent that we are now clinging to what remains of political common ground with our fingernails. Social media is a toxin coursing through the veins of the body politic with such deleterious effect that the most sober historians and political operatives are sincerely wondering if the Republic can survive another decade. It ain’t copyright enforcement that sends QAnon wackos to Congress.

Yet, to the tech-utopian, any effort to allow copyright owners to better protect their works online will unavoidably, and unacceptably, silence someone’s speech somewhere. In fairness, this is true. It is inevitable at times and must be remedied on a case-by-cases basis. Further, I see no reason why intentional abuse of DMCA to silence speech (e.g. criticism) cannot be more strongly proscribed through statutory reform if need be.

But citing “speech” as a generalized framework for debate is too broad and has little to show for itself as a social benefit to date. Aside from the fact that speech is silenced every minute online through many modes (e.g. bullying or platform moderation), there is no way that anyone can measure how much speech is currently silenced, or how much more or less would be silenced by improving the DMCA for rights holders. It’s counting grains of sand in the desert.

Ironically enough, Twitter announced over the weekend that it would delete tweets by anyone hoping the president dies from COVID-19. And while there are several reasons why this is sound policy for Twitter, it happens to be one of the few occasions when a platform would censor a prime example of protected speech. And, as one commenter rightly pointed out, Twitter has left intact volumes of missives hoping for the sexual assaults and deaths of women who speak out on various issues, including actual threats that transgress any claim to the speech right. So, we should dial down the speech rhetoric until it describes what the world actually looks like, not that Barlowian “home of mind” that never existed.

Into this long and repetitive debate, I think a fair market summary of the DMCA’s status quo is as follows: The major copyright owners enforce their rights through the use of some technological measures and the notice-and-takedown system, albeit with a ceaseless, dynamic, and expensive process that has little effect addressing the volume and rate of infringement. The small rightsholders barely enforce their rights at all through notice-and-takedown and generally give up trying. The user-generated platforms continue to profit substantially from third-party infringements against both small and large creators. And the 229-million of us Americans using the web comprise billions of transactions every day that have nothing to do with copyright.  

Google v. Oracle X: The Tao of Google

While reading a few other articles about this case, articles written by actual legal experts, I was reminded that Google v. Oracle, despite its epic scale and likely significance as a precedential ruling, is, in fact, not terribly complicated. At least it shouldn’t be. What has made the case complicated of course is Google’s obfuscation in an attempt to win. But the problem for creators, both within and beyond the software industry, is that Google’s key claims in its defense run afoul of some core copyright principles.

Keep in mind that the only reason Google declined to license the code at issue, as many other commercial developers have done, was its refusal to allow interoperability with the rest of Java. It was this aspect of the license agreement that presented a hurdle to Google’s desire to monopolize as much of the mobile market as possible, a goal the company has since achieved. Thus, Google’s defenses may be viewed skeptically, as legal constructs erected after a strategic decision to infringe and are, therefore, rather circular and distracting in their reasoning. But such is the Tao of Google.

The Simple Story:  Verbatim Copying & Copyrightability

Google copied over 11,000 lines of Java code. This fact is undisputed. Also undisputed is the fact that computer code is expressly protected by copyright as of at least 1980, and that unlicensed copying of computer code, therefore, is an infringement of copyright. But says Google, the kind of code they copied (what the industry calls “declaring code”) should not be protected by copyright. However, another undisputed fact is that while adding computer code to the copyright law, Congress was confronted with the very question Google raises and determined that code is code with respect to its protection as a “writing.” Nevertheless, Google persists …

According to those who know, Java is very good code. Consequently, a lot of programmers are fluent in the use of Java. And that is the reason Google copied what it did—with the aim of attracting developers to create apps for the Android platform as quickly as possible. While that is an understandable market interest, it does not alleviate the obligation to license the work at issue.

Yet, Google seeks to conflate its exigent “need” to copy with the exception under copyright law known as the “merger doctrine.” Merger applies when there is only one way, or very few ways, to express a particular idea—or, as is the case with software, to perform a given set of functions. Google hopes that the Court will overlook the undisputed fact that there is more than one way to write code to achieve the same functions for which it appropriated Oracle’s work—that in fact other developers have done so. Google’s argument asks the Court to apply the merger exception after a work attains popularity, where it would not apply at the instant of authorship …

The Simple Story:  Transformative Fair Use Actually Requires a Transformative Purpose

Should Google fail to convince the Court that 1) the code at issue is not code under copyright law; or 2) that the code at issue falls under the merger doctrine by virtue of its popularity, Google naturally asserts a fair use defense. This is one reason Google v. Oracle is a very big deal. As former Register of Copyrights Maria Pallante notes in an interview with Copyright Alliance CEO Keith Kupferschmid, the Supreme Court does not take copyright cases very often and that this will be the first time in 26 years that the justices will consider a fair use defense. Rightsholders everywhere should hope that the Court holds that fair use in the 21st century, despite all the shiny new objects, is not any different than it was in the 20th century.

Especially with regard to “transformativeness,” lately asserted in just about every alleged infringer’s defense, the Supreme Court’s opinion about Google’s appeal to transformativeness in this case may have profound effect on copyright enforcement for decades to come. Although Google used the code at issue for the same purpose for which it was written, it seeks to obfuscate this inherently non-transformative use behind a cloud that the mobile market itself has been, in a sense, transformed. Notwithstanding the fact that Apple, not Google, revolutionized the landscape of mobile …

This may be the most hazardous argument in Google’s quiver for copyright owners in every category. It alludes to the fourth factor in the fair use analysis, which weighs the potential market harm to the original work. This consideration is meant to look beyond the immediate harm that may be caused by the contested use itself. It asks what would happen to the market for the original if the same conduct were widespread and engaged in by multiple parties, and it asks whether the alleged infringement has appropriated the original author’s right to prepare derivative works. In Google v. Oracle, the answers to all of these questions should strongly disfavor a finding of fair use under factor four.

If the Court were to agree that the code at issue is not copyrightable, that would have major implications for the software industry and, in the view of many, be a misread of the 1976 Copyright Act and Congress’s 1980 amendment to it. Alternatively, if the Court finds that the copied code is correctly protected by copyright, but that under the fourth factor in particular, Google’s use was a fair use, the ruling could be disastrous for creators in every medium. It would say to every creator that whoever has the resources to attain market share the fastest is justified in appropriating any work to serve that purpose. That is anathema to copyright’s purpose.

It is no accident that this consideration is so prominent in a case involving Google. Almost nobody can compete with that company when it looks to dominate a segment of the digital marketplace. And copyright law remains one of the few bulwarks independent creators (and society in general for that matter) still have against Google’s near-monopoly power. Why else have they spent so much time, money, and energy trying to weaken copyright, along with anti-trust and privacy protections? Because the Tao of Google is no joke.

DMCA Hearing 5: More Consensus Than You Might Think on Section 1201

On September 16th, the IP Subcommittee of the Senate Judiciary Committee held hearing number five in its ongoing review of the 22-year-old Digital Millennium Copyright Act. The subject was Title I of the DMCA, also referred to as §1201, which proscribes the circumvention of Digital Rights Management (DRM) technologies used to protect copyrighted works distributed through digital systems. When the DMCA was enacted in 1998, as the market was shifting away from the purchase of physical copies of works, DRM was predominantly used to control licensed access to creative content (e.g. eBooks) that falls under traditional categories of copyright law.

With regard to the ordinary consumption of music, movies, TV shows, books, etc., §1201’s protections have been highly successful for producers and consumers. Critics who scorn DRM on principle (i.e. hate copyright) are unlikely to acknowledge that all the variety and instant access we enjoy—from Audible to Netflix—is entirely founded on DMCA’s anti-circumvention provisions. At the same time, readers are probably familiar with, and sympathetic to, complaints about DRM preventing farmers from repairing tractors, patients from accessing medical data, security researchers from analyzing vital systems, or the blind and deaf from accessing reading material. Yet, despite a considerable volume of headlines on these issues, last week’s testimony sounded more like consensus than conflict.

Professor Blake Reid, Director of the Samuelson-Glushko Technology Law & Policy Clinic, testified on behalf of various constituencies that have been unintentionally harmed by §1201 and addressed, for instance, the matter of access to electronic books for the blind and deaf. Although organizations representing blind and deaf readers have successfully petitioned the Copyright Office for §1201 exemptions in its triennial rulemaking process, Reid and others assert that the requirement to re-apply every three years is a high burden for these constituencies, merely to maintain exemptions that have no opposition.

Not surprisingly, the Copyright Office seems to agree with Reid on this point. In her testimony before the committee, General Counsel Regan Smith stated that because the statute does not give the Office authority to grant permanent exemptions, Congress could make the process more efficient and fair by amending §1201 to explicitly and permanently exempt deaf and blind readers, along with other frequent petitioners (e.g. security researchers), who have consistently been granted exemptions without opposition. It stands to reason that the Copyright Office would prefer not to spend its limited resources reviewing and re-reviewing the petitions of these parties, to say nothing of the fact that many interests in circumventing DRM often fall under the jurisdictions of other agencies, like the EPA, more than the Library of Congress.

The apparent consensus between the Copyright Office and many of the stakeholders represented by Professor Reid suggest to me that all those headlines, which have fed a lot of §1201 rancor over the years, may be outdated with respect to the task before the IP subcommittee today. At least in the consideration of various categories of petitioners, there appears to be ample room for negotiating a compromise bill calling for a wider spectrum of permanent exemptions.

Where Conflict Persists

More generally, however, there remains an ideological split over the nature of §1201. Reid and others view the prohibition against circumvention of DRM as a prior restraint on those parties who would only circumvent in order to engage in otherwise legal conduct. Thus, these critics of Title I would like Congress to amend the statute to affirmatively state that circumvention for non-infringing purposes is not a violation of law. The Copyright Office does not endorse this nexus between circumvention and infringement. Especially with regard to §1201’s prohibition against trafficking in devices used for circumvention, this nexus would undermine the fundamental purpose of the statute, which is the aforementioned fostering of eBooks, streaming platforms, and all those services we enjoy.

Personally, I am not unsympathetic to Reid’s view, at least in principle. If a documentary filmmaker uses software to rip a scene from a DVD to incorporate into her film in a manner that would be considered a fair use, the act of circumvention does seem incidental to her otherwise legal conduct, especially because the DRM is there to protect copyright, which, in this hypothetical, is not being infringed. So, I get where Reid et al are coming from when they argue that the act of circumvention itself should not be a violation.

On the other hand, these types of one-off examples are at least a little bit flimsy as practical matters. If the doc filmmaker uses AV clips in a manner consistent with fair use, the owner of those clips is highly unlikely to file a viable legal complaint. And absent a solid copyright infringement claim, the prospect that the rightsholder will file suit solely on an allegation of circumvention seems like a dubious path for most attorneys to follow.

Doubtless, Reid and I could play Pong with different examples tilting toward one hypothetical outcome or another. But it seems to me that the more exemplary users represent a class (e.g. teachers), the more they belong in the above discussion about candidates who deserve permanent exemption, rather than serve as a rationale to codify an infringement nexus. In fact, both teachers and documentary filmmakers have been granted limited exemptions in prior USCO rulemakings and are both viable classes for the committee’s further consideration.  

If I can presume to read between the lines of the testimony, the Copyright Office has an interest in resolving many of the unanticipated conflicts that have arisen from the protection of software integrated into devices and machines that are not typically within the scope of copyright law; but the Office also has a duty to §1201’s original purpose to secure the rights of copyright owners to control the manner in which their creative works are accessed. Endorsing an infringement nexus would perhaps resolve the former, but it would also betray that latter. Hence, a more nuanced approach to legislative adjustment should be sufficient to satisfy all parties with balanced results.

See full summary of the hearing at Copyright Alliance.