DMCA Section 512: A Relic of the Past, or An Engine Requiring Fine Tuning?
by Neil Turkewitz
Senator Tillis is spearheading a much needed review of the Digital Millennium Copyright Act (DMCA), and called for submissions from the public to inform the process. This follows the delivery of a comprehensive report from the Copyright Office in which the Register conveyed the Office’s view that the DMCA is not currently meeting the goals of Congress underlying its adoption — in particular, that the DMCA, as interpreted, is unbalanced to the disadvantage of the creative community: “The Copyright Office concludes that the operation of the section 512 safe harbor system today is unbalanced…The Report highlights areas where current implementation of section 512 is out of sync with Congress’ original intent, including: eligibility qualifications for the service provider safe harbors; repeat infringer policies; knowledge requirement standards; specificity within takedown notices; non-standard notice requirements; subpoenas; and injunctions.”
There will likely be many submissions from interested parties — both challenging and defending different aspects of the DMCA. I filed a submission on behalf of no one but myself, posing a very broad question: should we keep Section 512 of the DMCA at all? You may find my entire submission below.
Let me begin by thanking you for your attention to the importance of ensuring that America’s copyright laws continue to reflect the needs of American society — including in particular that, as the Supreme Court observed in Harper & Row v. Nation, it continues to serve as the “engine of free expression.” I submit these comments in my personal capacity; on behalf of no one but myself.
I was, as a representative of the Recording Industry Association of America (RIAA), directly involved in the negotiation of the WIPO “Internet Treaties” (the WCT and the WPPT), and in the development of the DMCA. As a long time advisor to the Office of the United States Trade Representative (USTR) and the Department of Commerce as a member — and ultimately Vice-Chairman, of ITAC 15, I also provided continuous guidance on the relationship of the DMCA to the national interest as reflected in US trade policy.
I will be brief in my comments — and perhaps somewhat controversial. At the end of this document, I attach some articles that I have written that address many of the questions that you have raised. I hope that you find them useful as you continue this critical examination of the DMCA.
There is only one observation that I will make. It would appear that no one likes DMCA Section 512 — with the possible exception of large online platform operators who testify in support of Section 512, even when they’re generally quite critical of it in the media. Creators don’t like it because it creates an impossible enforcement burden and exempts from responsibility the party most capable of taking action to address infringement. Could you imagine asking law enforcement to address only the demand/consumption side of an illegal trade, not dealing at all with supply/distribution? Of course not — no one would think that a solution should focus on the ants, and not the ant farm. Yet that is precisely how Section 512 works.
Users are no more fond of Section 512 than creators, feeling that there are far too many takedowns of non-infringing content resulting in a culture of censorship. They see takedowns, and Section 512 in general, as a sword used too easily by what they see as rapacious copyright owners.
That leads me to what I believe is a logical and sound solution — I think you should consider repeal Section 512 in its entirety. Get rid of what too many see as a tool of control wielded by parochial interests. Get rid of a mechanism that suspends traditional tort principles of liability. DMCA’s Section 512 was adopted at a time when the “internet” was seen as existing in a plane unto itself, removed from the “real world” and capable of being regulated (or not-regulated) in a sui generis manner. The past two decades have, I believe, demonstrated the fault in this premise, and a growing recognition that online is very much a part of “the real world.” In particular, the suspension of common law principles of responsibility and the forced halt in the development of common law rules of liability which had been taking place prior to the adoption of the DMCA, has greatly harmed the ability to fully develop the online marketplace for works protected by copyright, to the great harm of creators and users alike.
In the absence of Section 512, platforms would have incentives to develop and implement structural systems to deter infringement that will be much more effective, and less likely to be perceived as heavy-handed, than mere reliance on notice and takedown. The world is full of infinite opportunities fueled by technological capacity, yet we are locked in a finite dance of notice and takedown without noticing that everything has changed around us. In many respects, notice and takedown has turned us into North-Going Zaks and South-Going Zaks. Meanwhile, the greenery to the east and west remains wholly unrealized.
The potential of technology to explore near infinite opportunities was perfectly on display in a recent YouTube project, fittingly called “An Infinite Music Video.” As reported by Rolling Stone:
“YouTube is releasing what the company calls “the world’s first infinite video,” which takes short clips from tens of thousands of “Bad Guy” covers and randomly links them together, creating a new visual every time. Thanks to AI that can recognize song structure, the clips are always strung together to fit the sequence of the song.
To create this experience, Google Creative Lab had to perfectly align the audio from thousands of videos with different tempos, instruments, keys, and styles,” YouTube said in a blog post Monday morning. “Conventional algorithms couldn’t do the trick alone, so we turned to machine learning. For videos that deviated from the original track — like acoustic or acapella versions — we built a neural network that could predict matches between covers and the original. With this data, we were able to line up all kinds of different covers and switch seamlessly between them.”
My immediate reaction was “Cool, they can seamlessly sync up all of these disparate threads by building a bespoke neural network, but they can’t build a better technology to detect and deter infringement at the source?” Can you imagine if this kind of focus and creativity — the power of technology, was unleashed at resolving the problem of piracy? Notice and takedown locks us into a model of commerce akin to driving a Model-T in a world of Tesla’s. Congress shouldn’t play at the edges, trying to marginally improve the performance of a used car that’s already outlasted its usefulness. We need a paradigm change, not a tune-up.
In a world without Section 512, platforms would — by virtue of being potentially liable under theories of vicarious liability and contributory infringement for failure to use reasonable care, have incentives to develop technologies at the systems level that would be far more effective at addressing infringement and creating the kind of economic opportunities that have eluded creators in the two decades since adoption of the DMCA.
This was what Congress had in mind in adopting Section 512 — that technologies would be employed at the systems level to detect and prevent infringement, and that notice and takedown would be employed for materials that made it through the cracks, or relied upon by platforms on which infringement was uncommon. Instead — especially in light of judicial interpretations which effectively eliminated red-flag knowledge in favor of actual knowledge, notice and takedown became the de facto standard for online enforcement. It wasn’t built for that. Then, as a result of being forced to rely on a system of notices, those copyright owners that had the means to do so ramped up their notification schemes on the correct premise that the response to harm must seek to reach the scale of the harm. This of course requires a level of automation that increases error rates. As such, it is fair to observe that takedowns of non-infringing materials are a direct result of forced over-reliance on notice and takedown to address harm. Eliminating such reliance therefore has manifest and direct benefits for both the creator and user communities.
I also highlight that repealing Section 512 is not nearly as radical a proposition as it may first appear. Indeed, it is interesting and instructive to note that just two years prior to the adoption of the DMCA, Congress recognized in the Communications Decency Act (CDA) that the immunity granted to platforms under Section 230 should not extend to intellectual property. See: “Nothing in this section shall be construed to limit or expand any law pertaining to intellectual property.”
Why did Congress not extend platform immunity with respect tointellectual property violations? Because Congress understood in 1996 that relieving platforms of the responsibility for addressing copyright infringement did not serve the national interest, and that established legal doctrines were capable of providing appropriate guidance on the legal obligations of service providers. Doctrines of vicarious liability and contributory infringement articulated rules by which the party in the best position to address harm was not relieved of its duties to do so.
The Copyright Office neatly summarized these doctrines in its 2004 statement before the Senate Judiciary Committee in relation to Congressional consideration of “The Intentional Inducement of Copyright Infringements Act (INDUCE).”
The Copyright Office wrote:
“Vicarious liability requires two elements: (1) the right and ability to supervise or control the infringing activity; and (2) a direct financial benefit from that activity. It is closely related to the doctrines of enterprise liability and respondeat superior in tort law.
There is another form of secondary liability in copyright law, “contributory infringement,” which stretches back to 1911. As the Second Circuit Court of Appeals has explained, contributory infringement occurs where “[o]ne who, with knowledge of the infringing activity, induces, causes, or materially contributes to the infringing conduct of another.” In general, the two elements of contributory infringement are (1) knowledge of the infringing activity; and (2) material contribution to the activity.”
The application of these principles was well-articulated in the 1996 decision of the 9th Circuit in Fonovisa v. Cherry Auction in which the Court set out a framework for considering vicarious liability and contributory infringement. The 9th Circuit held, inter alia:
“In Gershwin, the defendant lacked the formal, contractual ability to control the direct infringer. Nevertheless, because of defendant’s “pervasive participation in the formation and direction” of the direct infringers, including promoting them (i.e. creating an audience for them), the court found that defendants were in a position to police the direct infringers and held that the control element was satisfied. 443 F.2d at 1163. As the promoter and organizer of the swap meet, Cherry Auction wields the same level of control over the direct infringers as did the Gershwin defendant. See also Polygram, 855 F.Supp. at 1329 (finding that the control requirement was satisfied because the defendant (1) could control the direct infringers through its rules and regulations; (2) policed its booths to make sure the regulations were followed; and (3) promoted the show in which direct infringers participated.”
Senator Tillis, as you have already learned, including from the comprehensive report delivered by the Copyright Office at your request, the original Congressional aim of the DMCA has failed to materialize. Provisions designed to promote reasonable accountability have instead been interpreted as eliminating red flag knowledge, forcing reliance on an over-burdened notice and takedown system not capable of addressing infringement at the scale at which it proliferates — resulting in takedowns that are both under and over-inclusive. The promise of the DMCA largely hinged on the development and use of technologies that would be accommodated by all internet actors to enhance legitimate commerce by reducing the unfair competition posed by infringement. Those technologies failed to materialize because platforms had no direct financial interest in developing such robust systems, and because the DMCA provided immunity regardless of whether they were developed and used or not.
You have an opportunity to create such an incentive and to unlock the potential of digital technologies to drive a new cultural renaissance while simultaneously improving the user experience. As you yourself have observed: “Rather, than tinker around the edges of existing provisions, I believe Congress should reform copyright law’s framework to better encourage the creation of copyrightable works and to protect users and consumers making lawful uses of copyrighted goods and software-enabled products, respectively.” You can do that by repealing Section 512. At a minimum, it seems like it would be a good exercise to examine why, if at all, a legal environment without Section 512 is inadequate, rather than proceeding with a failed status quo as the default option. The starting point for the journey affects the process, and ultimately the result. This is a time for bold and decisive action that changes the underlying paradigm — to encourage responsible practices rather than to reward indifference. I thank you for your leadership.
Respectfully submitted,
Neil Turkewitz
Attachments:
Addressing the Failure to Develop STM’s
An Examination of Perspectives on the Digital Marketplace
& this.
An Independent Creator’s Experience with Notice & Takedown
A Look at Fraudulent Notices & Counter-Notices
A View of the Implications of Article 17 of the EU Copyright Directive
How safe harbors have undermined “market” conditions for music creators
