Category: Copyright

Implications of YouTube’s Copyright Match System

Last month, the European Union voted against key copyright enforcement provisions as part of its Digital Single Market initiative. Specifically, the proposal known as Article 13 called for the 28 member states to work with multiple stakeholders to develop and implement filtering technology that would, in theory, prevent unlicensed, copyrighted works from being uploaded onto user-content-supported platforms.

Article 13 was labeled by its opponents as a call for “censorship machines,” and as usual, the refrain was shouted from the rooftops that adopting any such filtering would lead to the end of free speech online and destroy “all that is good and pure” about the internet. To be sure, the tone and methodology of the campaign against these provisions reeked of Silicon Valley money and tactics, but whether you believe that or not, one talking point among critics stands out, which inspired this post: that the EU’s call for filtering would harm new creators.

Granted, “creator” in the rhetoric of groups like EFF includes literally anyone who posts anything online; but if we limit our data to that paragon of new creators—the successful YouTuber—YouTube itself made an interesting announcement almost concurrent with the defeat of Article 13. The platform launched its Copyright Match tool to protect YouTubers against unauthorized re-posting (freebooting) by other YouTubers.

Traditional rights holders have earned this moment of schadenfreude after being lectured to for years to get on the future bus and quit whining about their copyrights. They should follow the example of “new creators” working in “new models” that “bypass gatekeepers” and obviate the need for copyrights. Of course, it was inevitable that as YouTubers became entrepreneurs, they would feel entitled to the revenue from their labor (as they should) and that YouTube would have a vested interest in protecting the copyrights of its profitable video-makers—at least from other video-makers.

Using technical measures one might call “filters” (or dare we say “censorship machines”?), the new Copyright Match system works by identifying the first upload of a new video and associating that file with the presumptive owner of the work. Then, if and when matching videos are uploaded to YouTube, the original creator is notified and given the option to do nothing, to ask YouTube to remove the Match, or to get in touch with the uploader of the Match.

Hypocrisy Much?

If this sounds to the experienced observer like an intramural version of a take down/stay down provision achieved through technical measures, that’s because it is. And experienced observers probably remember that all previous proposals for take down/stay down, whether statutory or technological, have been labeled by industry-funded “activists” as internet-killing initiatives. In fact, during the May 2016 hearings about the DMCA, one of the dumb-but-effective talking points was that any mandate for such technical measures would “entrench” the market dominance of YouTube. (Yes, laughing through tears is the right response here.)

Traditional rights holders who have spent hours of their lives trying to identify and stop unlicensed uses of their works on market-dominating YouTube will quickly recognize the duplicity in launching Copyright Match. “Why should only their ‘chosen’ get access?” asks Grammy-winning composer Maria Schneider, one of many artists who will attest to the opaque and labyrinthine Content ID system rights holders theoretically use to track and control use of their works on YouTube.

What is not generally understood is that even getting access to Content ID varies wildly depending on a rights holder’s relative presence on YouTube and his/her interest in monetizing unauthorized uses vs. taking down unauthorized uses. Guess which one YouTube favors. Again, Schneider explains …

“YouTube always says that independents like me, to whom they’ve denied ContentID, can get access to the same tool via a third party. But what they don’t admit publicly is that this is only possible if we’re willing to monetize at least some of our work. So, independents like me, who want no part of monetization and simply want to block illegal uses of our works are just out of luck. And I might add that this technology we’re wanting to access has actually been around for twenty years—longer than YouTube has existed!”

I spoke to an independent artist, who prefers to remain anonymous for fear of retaliation by YouTube’s Content ID group.  He does have Content ID and acknowledges that he’s probably a “thorn” in the side of YouTube’s Copyright Department team because he actively employs the system only to stop unlicensed uses of his compositions and sound recordings. And lest anyone think he’s responding to “new creators” making possible fair uses of his music, his most important Content ID-related takedowns have been aimed at global brands and Fortune 500 companies using his music for marketing purposes without a license.

It’s worth noting that the fact that this artist chooses to remain anonymous due to concern that YouTube would delete his Content ID account out of spite speaks volumes against every claim of tech-utopian bullshit Silicon Valley and its network of EFFing dissemblers have been slinging for years. As David Lowery explained in 2016, YouTube is a monopsony, a market with a single buyer, which means they get to make, break, and change the rules as often as they like, and the “sellers” can just eat it.

In this regard, it will be interesting to see if Content Match leads to disputes YouTuber-to-YouTuber and how the company will handle these, if it does. For instance, it is not clear at this point that YouTubers whose uploads are wrongly identified as “Matches” will have any kind of counter-notice remedy available to them.

Although the company’s video explaining the new system urges YouTubers requesting Match takedowns to “consider fair use,” it will be truly fascinating to see whether YouTube gives a damn about fair use among its own microcosm of creators. For sure, general users of the platforms have never been effectively dissuaded from uploading a wide range of files that could never qualify as fair uses.

None of this should be taken as a dig against YouTubers. To the contrary, I think many of them are brilliant artists and deserve to protect their interests and rights as much as any other creator. But this apparent initiative to protect their interests points to another aspect of YouTube’s ever-changing relationship to copyright enforcement and its relevance to the fight over Article 13.

Don’t Let the Internet Become YouTube?

Not that long ago, YouTube was consistently cited as the apotheosis of the utopian belief that the web will empower creators without gatekeepers—and without copyrights. But where this Copyright Match announcement becomes intertwined with the campaign against Article 13 is that some pundits against the proposal lately cite YouTube as a cautionary tale—asserting that the platform’s often-inconsistent application of copyright protection policies and technical measures is exemplary of what should not be done internet-wide pursuant to Article 13. The claim appears to be that because YouTube’s Content ID system has allegedly fostered rampant false strikes, resulting in unfair channel deletions, this generalized stifling is what the “entire internet would look like” if the EU moved forward with the kind of filtering proposed.

While there is certainly anecdotal evidence—some of it compelling—of Content ID error and abuse leading to improper strikes on YouTube, I have yet to see any evidence to support the claim that this problem is both rampant and increasing across the platform. As is often the case, activist groups or observers who have no skin in the game tend to exaggerate anecdotal evidence into statistical assumptions. Or as our anonymous artist puts it, “In 100% of the anti-Content ID statements I’ve ever heard over the years, 100% of the complainers had 0% vested interest in the system: they’re either Google-funded anti-copyright groups or individuals on some kind of personal crusade.”

In this creator’s direct experience with false identifications, he notes that “With about 100,000 Content ID claims in my dashboard since late 2012, I can say that YouTube has delivered me a mistaken ID about 10 times. The anti-copyright crowd will take that as evidence that the system needs to be dismantled or destroyed. I just don’t get it. The perfect shouldn’t be the enemy of the good.” And that’s for creators, who can avail themselves of Content ID, which does not include the creators identified by Maria Schneider who do not have access to any remedy via Content ID.

Competing Narratives

So, in context to the proposal that Article 13 filtering would “stifle new creators,” we have at least three narratives that compete and crisscross in ways that can be hard to track, if you’re not directly engaged with these systems. First, because Copyright Match is a response to YouTuber complaints about freebooting, it reveals that “new creators” don’t like copyright infringement when it happens to them (ergo copyright is not obsolete). Second, Copyright Match implies that filtering technology of this nature can be implemented without destroying a whole platform or stifling new creators. And third, Copyright Match is at least indicative of technology that could help non-YouTuber creators enforce their rights, but it will not be made available to them because it isn’t in YouTube’s interest to do so.

One thing the introduction of Copyright Match illustrates for sure is that creators are creators—whether traditional or new, they feel a sense of ownership in the products of their labor. And from this premise comes the foundation of copyright and systems for protection that will begin to make “new models” look a little more like “old models.” It’s what happens every time a business discovers it is codependent with talented people.

Copyright Found After Brief Search

Well, artists and authors, I guess you can pack it in. Professor Glynn S. Lunney, Jr. of Texas A&M School of Law has declared copyright dead in a recent 12-page paper that is presumably a digest of his new book Copyright’s Excess: Money and Music in the US Recording Industry. Apparently, Lunney first announced copyright’s demise in a 2001 article tautologically named Death of Copyright, in which he then identified the proximate cause of death as the DMCA. But now, Professor Lunney states, “Copyright is dead. But it was not the DMCA that killed it. It has become increasingly clear to me that copyright, as a law that serves the public interest, was only ever a dream. In the real world, it never existed at all.”

So, copyright didn’t die. It’s been a ghost all along. In his new paper Copyright Lost, Lunney argues that the law’s fundamental promise is “more money equals more works,” and that because his data reveal this to be untrue—in fact, he finds the opposite to be true—copyright has never been the public-serving legal framework we’ve been taught to believe it is. But I think the real phantom Lunney is pointing to is his own scholarship, which begins with a straw man and then proceeds to cite irrelevant (if even true) data to reject a false premise.

Copyright is Not Synonymous with Money

Copyright’s whole purpose does not boil down to “more money equals more works.” The legal framework itself is agnostic with regard to how much money an individual author, or group of authors, actually earns. Yes, the bundle of rights opens up a range of possible revenue streams, but these are uniform (or constant) for the author who earns a lot of money and the author who earns very little.

For instance, the copyrights on a James Patterson thriller are identical to those on Lunney’s new book (I know, right?), even if Lunney’s contract with Cambridge University Press is quite different from Patterson’s contract with Random House. And the copyrights underlying those contracts have no bearing on the market reality that Patterson will outsell Lunney at a ratio of a gazillion to one. Money is a factor in Patterson’s staggering output, but the copyright itself is not predictive of that output.

It is true that one argument for copyright’s value is that the royalties from precedent works can provide authors with the resources to create new works, but this is not the sole purpose of copyright, and it is more than a little mercenary for Lunney to propose otherwise. The author who only produces a single work, even if that work is wildly profitable, does not “owe” society more works because society “granted” the copyright in the one work.

Authors and artists are mortal beings, some gifted with the right combination of skills, timing, luck, and even frailties, that make them especially productive. Others struggle for decades to produce a modest body of work. And, of course, there are external factors like family, health, world events, etc. beyond the author’s control that may affect her output of work. In this regard, Lunney is outright insulting to creators when he invokes a principle of labor economics that may be true among certain classes of workers but is rarely true among artists. He writes …

“Once wages have increased to the inflection point, individuals are earning enough that they want to buy more leisure rather than work. As a result, the income effect begins to outweigh the substitution effect, and the labor supply curve starts to bend backward. Beyond that point, further wage increases will lead the individual to work less, rather than more. In the late 1990s, copyright ensured an effective “wage” for our top artists and authors far in excess of their reservation price, and potentially above the point at which the labor supply curve began to bend backward. Thus, I argued that reducing revenue might actually lead some superstar artists to work more, rather than less.”

Right. Tell that to David Bowie working on Blackstar with his last dying breath. Lunney’s appeal to this principle may accurately describe a wide range of employees who would not necessarily work if they didn’t have to, but it is a minority of creators who actually stop working, no matter how much or little they earn. What academics consistently misunderstand in these analyses is that unlike most people who work to live, most artists live to work.  You simply cannot compare the pecuniary motivations of the average creator with those of some executive at a bank who would “rather be sailing.” Jimmy Buffett sails and has earned a few bucks, but he’s never stopped making music.

Sketchy Data Supporting a Dubious Theory

Nevertheless, having established an incomplete, if not outright false, premise for the purpose of copyright, Lunney sets out in search of evidence to support his theory that higher revenue consistently fails to yield “more and better” works. And unsurprisingly, he finds exactly what he’s looking for in data that not only fails to prove his point, but actually has little to do with copyright. Lunney’s theory can be summarized as follows:

The creation of the sound recording right increased revenue from sales and licensing of recorded music, but contrary to expectations, high-revenue periods were not high-output periods. In fact high output occurred during low-revenue periods.

To arrive at his conclusion, Lunney poses a broad question and a specific question. Broadly, he asks whether “more and better” music was produced at a lower rate after the phenomenon of mass piracy (Napster), and more specifically, he asks whether periods of peak revenue correspond to periods of “more and better” output. He defines “more and better” as “music people want to listen to,” which is a dubious metric because it does not inherently point to quantity and diversity. If there were only 100 songs in the world, they would be very popular indeed.

To observe “what people want to listen to,” Lunney cites Spotify data and finds, not surprisingly, that there is a sharply ascending peak in streams of music produced between 2000 and 2010, and he also shows that peak revenue periods like the 1990s correspond to a substantial dip in current popularity among Spotify listeners. From these data, Lunney concludes that we can “reject the notion that more revenue led to more and better music.”

No, I don’t think we can draw this conclusion, given the wide range of possible conditions that must be considered to responsibly assess the data. These include, among others, the fact that the most popular music at any time is usually the newest music; that more than 34% of Spotify users are under the age of 35; that Spotify streaming data is very narrow relative to Lunney’s broad inquiry; that some of the best niche music to emerge in the 1990s might be accessed by fans via other media; and that the concept “better” is far too subjective to simply sweep into the bucket of contemporary popularity and call it evidence.

For instance, Lunney’s Spotify data reveal a noticeable peak in the mid 1980s. If he asserts that the adjacent dip in the 1990s is literally caused by a corresponding rise in revenue for that decade, is it truly reasonable to assume that a corresponding “dearth” in revenue is the most rational explanation for the popularity of 80s music on Spotify today?

Or does this peak simply reveal that the next largest demographic of Spotify users is my generation, who grew up on 80s music? Or perhaps it’s explained by the fact that a lot of our kids also like 80s music because we’ve exposed them to it? Or might it have something to do with fact that a lot of new musical works are drawing inspiration from 80s sounds right now?  What happens to Lunney’s theory if 90s music becomes more popular in five years because trends are cyclical?

I don’t know these answers, of course, but because Lunney is measuring contemporary music taste (and only on Spotify), his attempt to explain current trends by virtue of previous years’ revenue alone rings a little hollow as an economic analysis, to say nothing of citing his findings as proof of copyright’s worthlessness.

“It is no coincidence,” Lunney writes, “that the most prolific artists in the study, the Beatles and Taylor Swift, had their first 100 hits in the low revenue years of 1964 and 2006, respectively.” It isn’t a coincidence? How does Lunney know this without accounting for, I don’t know, about a thousand other variables? More to the point, how does he sincerely invoke the Beatles—the Beatles!—in a paper which asserts that as income increases, artistic output decreases?  What does the overall industry revenue of 1964 have to do with the fact that two years later, the Beatles heard Pet Sounds and then produced Sgt. Pepper’s in 1967, which went on to sell over 30 million copies?

Maybe Lunney means some other Beatles and not the guys who started out playing cover songs and dance music in clubs then, when they earned real money, produced some of the most revolutionary sound recordings of all time. Those Beatles, Lunney would argue, decreased their output as they attained wealth from album sales. McCartney and Starr are still making music. And does anyone seriously doubt that Harrison, and especially Lennon, would still be writing and recording, if they were alive?

Little To Do With Copyright

Most importantly, even if Lunney can support his theory that lower revenue periods produce “more and better” music, he is fundamentally incorrect to associate those findings with the sound recording copyright passed in the U.S. in 1972. Lunney states, “For the recording industry over the last fifty years, when copyright protection was strong and effective, it forced consumers to pay more for music.”

That is simply a false statement. Because other than comparisons to outright piracy, the sound recording right has never had much to do with retail prices for recorded music. Prices of albums, and then CDs, rose along with the price of other consumer goods, and whatever cultural excesses one might ascribe to the recording industry of the 1990s was also true of business in general in that decade. Greed was evident in every sector.

But the sound recording right itself is largely a B2B protection, implicated when one artist samples a recording made by another artist; when a sound recording is synced to a motion picture or TV show; or when a sound recording is publicly performed via certain digital platforms. The sound recording right has never played much of a role in consumer sales of music for the simple fact that consumers had neither practical alternatives to—nor major complaints about—purchasing physical copies of albums.

Even in the digital market, the sound recording protection underlies a limited set of exclusive rights that remain largely a matter of licensing among business entities and rightholders, and these have even less to do with consumer “prices,” which now range between zero and about $120/year for access to 25-million tracks on Spotify (if we’re sticking to Lunney’s limited data).

Finally, it must be noted that Lunney has focused all of his attention—worthy or not—on the music industry and the sound recording copyright post 1972, and this is a rather slim platform from which to declare “copyright is dead.” Surely, Professor Lunney is aware that copyright encompasses far more than sound recordings and that perhaps his diagnosis of the whole body of law is at least premature, if not outright quackery. I suspect creators will insist upon a second opinion.

New Paper Takes on the Myth of Free

Zeno’s Paradox describes physical change as an illusion. Zeno of Elea, in the 5th Century BCE, postulated that in order to travel any distance, one had to first travel half that distance, and before that half could be traversed, one had to travel half of the first half, and so on. And because space could be infinitely divided, traveling through space would seem mathematically impossible but for the fact that anyone running headlong toward a brick wall will suddenly realize why we call Zeno’s theory a “paradox.” Or as Professor John M. Newman might put it, apropos his recent paper The Myth of Free, the sudden impact with the wall will be an encounter with the “friction of the real.”

Even a casual observer of the real world can reasonably conclude that we are not heading toward a market utopia in which goods and services are both free and abundant. Yet, despite the empirical evidence all around us, Newman’s new paper in the George Washington Law Review asserts that the Myth of Free is not merely pervasive among academic theorists but has been wreaking havoc in applied law and policy for years. He writes …

“Legal institutions have already begun to grant an undeserved protected status to the suppliers of Free products. Such suppliers have received de facto or de jure immunity from certain laws, as well as favorable treatment in close cases. This is so despite the reality that these firms engage in for-profit transactions that are structurally identical to traditional, positive-price transactions. That unjustified protected status undermines the rule of law and distorts natural market competition, harming total welfare.”

Newman’s paper should be required reading by jurists and policy-makers seeking to understand, in economic terms, many of the same principles that artists and creators—the first victims of Free—have been saying anecdotally for quite some time.

Artists and Creators Know This Story Well

Day-to-day arguments about copyright are apt to be about specific points of conflict, like duration of terms, methods of enforcement, application of fair use, court opinions, etc. But the broader social, academic, and political campaign against copyright’s purpose—advocating changes from radical revision to outright erasure—is largely predicated on this economic belief that once the internet became publicly accessible, we crossed some threshold toward a future in which goods and services will inevitably be be both free and abundant. Where scarcity itself would be eradicated.

Newman’s paper is among the first, if not the first, to formally rebut the Myth of Free based on economic analysis. Covering more ground in his 74 pages than I can adequately summarize here, copyright owners and advocates will immediately identify with Newman’s critique of Free’s “major premise,” which asserts that zero (or near-zero) marginal cost drops the natural price of goods and services to zero (or near zero).

“According to the Myth, once information-based products could be reproduced and distributed digitally, the marginal cost of such products began to map onto the ever-halving cost of trendlines of processing, storage, and bandwidth. Eventually costs became so low that prices inevitably dropped to zero. Free was born.”

Authors of creative works will recognize this “marginal cost” premise as the head-banging gibberish they’ve encountered for years from people like Mike Masnick ‘splaining the alleged zero-cost of digital distribution of works while simply ignoring the high cost of development and production of works.

As Newman puts it, “…the Zero-Cost Premise fails to account for both the stubborn persistence of costs and what is referred to herein as ‘the friction of the real.’” To put that in practical terms creators know very well, even if the distribution cost of a movie really were zero, the cost to produce the movie is never going to approach anything like zero. But Newman doesn’t even need to go there, instead demonstrating that even digital distribution is never going to approach zero cost either …

“Digital processing, storage, and bandwidth do not occur in some abstract realm that transcends scarcity. They occur in the real world, which means they come with real costs. Those cost may be quite low, and may decrease lower still, but they cannot reach zero.”

So, even the premise underlying the premise of Free isn’t true, hence Newman’s reference to Zeno’s Paradox in this part of his paper to describe the fact that the cost of bandwidth itself will not continuously halve itself until it becomes “too cheap to meter,” let alone implicate that all goods and services will soon become both free and abundant as a logical extension of this tech-utopian fallacy. “Those who claim the existence of zero marginal costs fall into the same trap as those who chase the elusive dream of perpetual motion,” Newman writes.

Many of the footnotes in The Myth of Free read like a Who’s Who of “Freeconomics” hucksters—Mike Masnick, Chris Anderson, Mark Lemley, Jeremy Rifkin, Amanda Palmer et al, but I was particularly struck by the attention Newman gives to rebutting the premises of Professor Lemley in context to previous posts on this blog. Readers may remember that Lemley is the academic who has accused some of his colleagues—and by extension advocates like me—of endorsing a “faith-based IP” doctrine that he claims ignores empirical evidence. But as Newman makes quite clear in this paper, proponents of Free are engaged in magical thinking that goes beyond anything we might generously describe as faith and can more accurately describe as denial.

“…those who foresee a coming of age of abundance tend to miss—or misunderstand—a crucial element of the dynamic, long-run view: demand is a moving target. With that element in place, such techno-optimism is revealed as misguided. It is, as those who espouse the Myth of Free rightly recognize, a mistake to view technological innovation solely through the lens of short-run employment effects.”

In other words, when people like musical artists say the math of Free doesn’t add up, it’s not because they’re “clinging to old models,” trying to sell buggy-whips in a market of driverless cars.  It’s because the math of Free doesn’t add up.

As mentioned, it’s hard to imagine anyone observing the world today and believing that we are trending toward a post-scarcity utopia. But I suspect this particular idealism happened to emerge during a period when the failure to reign in certain excesses of capitalism demoralized the same generation that was being fed the hip and ebullient economics of Free. Quite simply, because the Napster generation collided with a market that was transparently bloated, corrupt, and offering dwindling opportunities, the conditions were probably ideal for evangelizing a vision of a post-scarcity future.  A vision of “sharing” that would somehow transcend market realities by virtue of “free” technology, which isn’t free at all.

Meanwhile, Newman describes in this TEDx talk that Free also has a specific neurological effect, which rewards gluttony and is consequently having a detrimental effect on both human activity and fundamental economics. “We as a society are dumping billions of dollars into developing ever more creative ways to first addict ourselves, then to extract from ourselves as much of our time and attention and personal privacy as possible, and then auction it off to the highest bidder.”

As if to underscore the point, Newman’s talk about the addictive nature of Free has just over 200 views, while Top 100 Viral Videos of 2107 has over 31 million.