Tag: Featured

Is “Machine Learning” Copying or Reading?

machine reading

I recently attended a round-table discussion on the subject of artificial intelligence and copyright.  The first of several engaging topics I thought warranted a post was the question of “machine learning,” which I put in quotes here with respect to one scholar who admonished against anthropomorphizing AI by using words for human activities to describe the actions of computers.  I think that view is fundamentally correct, though there is also grounds for analogy, as will be made clear by the following premise:

When you read a book, even if we might say, by way of analogy, that you are “copying” the content of that book onto your brain, this clearly does not infringe §106(1) of the copyright law proscribing unauthorized copying.  Since the author naturally hopes that you will read her book, such a prohibition would be absurd, even if you had an eidetic memory and could, if prompted, recite the entire work verbatim.  But if you used that gift to type from memory the entire book and made that document available, you would then violate more than one statute under the copyright law.

So, the question raised in regard to “machine learning” is whether the computer scientist who wishes to feed a corpus of books—say the anthology of American literature—into an AI should be required to obtain licenses for the works still under copyright.  Thus, the first analysis is whether the act of “copying” can be said to occur in this circumstance any more than it would be for the human reader who consumes the same body of literature.

It strikes me that if what the AI does in this case is ingest the corpus of books and almost instantly deconstructs those works by synthesizing them through a neural network, then the computer scientist has a pretty solid argument that no copying has taken place.   If the machine does not retain intact copies of works—or even large sections of works—-with the purpose of making those intact copies available to the human market, then this “machine reading” process is arguably analogous to the human whose reading does not infringe §106(1) of the copyright law.

That said, intent of the computer scientist may be a significant factor.  For instance, if the training of the AI will have a commercial purpose, this may suggest a requirement to license the works under copyright.  But intent can be very tricky on the leading edge of science because it is neither realistic, nor even desirable, to insist that every researcher know exactly where his experiments will lead.  This would nullify the process of discovery whence many great achievements have been made; hence, discovery is justification itself, and I suspect the tech companies would appeal to this rationale in regard to “machine learning.”

If the computer scientist’s goal is to see whether he can get his AI to “learn” about the American experience through literature, but he does not have a particular product or service in mind at the outset, it seems that copyright owners would be on fairly shaky ground to enjoin his use of the books.  As long as nothing that comes out the other end looks like any of the products that went in, it strikes me that this experiment exists beyond the statutory framework of copyright law.

Of course this portrait of the individual scientist beavering away in his modest lab to see what he may discover is not what is taking place in reality. We know perfectly well that major AI experimentation occurs in the R&D labs of companies like Google and Facebook, who are well shielded by trade-secret law from divulging what they are working on or for what purpose.  Like any other corporations, they are free to announce a new product or service without telling the public how they arrived at the latest result.

So, even if the use of copyrighted works as source material resulting in a commercial end might recommend some type of licensing regime, it may be very difficult to identify the threshold when the blind process of scientific discovery becomes a clear intent to exploit a commercial opportunity.  And, as mentioned, these companies would be under no obligation to divulge that eureka moment to anyone.  

On the other hand, the moment Google or Facebook did announce that new product, rightsholders could justifiably complain that a massive, highly-profitable corporation has used potentially billions of dollars worth of material without paying for any of it.  As one scholar at the round-table noted, tech companies may not use raw silicon for free, so why should they get to exploit millions of creative works for free, no matter what they’re turning that data into?

It’s a good question.  One that would seem to suggest a new subsection of the copyright law, and this would certainly be consistent with the fact that new forms of exploitation of works may demand equally new forms of compensation.  If nothing else, that type of statutory response could spare us all the tedious and false harangue that insists “copyright owners just want to stand in the way of innovation.”

That argument prevailed for far too long, and now the so-called innovators have a lot of splainin’ to do about their culture of blind disruption for the sake of disruption. Especially in light of the fact that AI may have some very profound effects on society as we know it, maybe this time around the copyright owners should be treated like experienced voices in the conversation rather than canaries wasting their breath in the proverbial coal mine.

Milton Never Said a Damn Thing About the Internet

I’ve been traveling and am, therefore, late to mention that the hotly-contested EU Copyright Directive passed last week. Not surprisingly, the usual critics have spared little hyperbole referring to the new legislation as the “end of the open internet” and a “disastrous decision.” Meanwhile, many of the copyright proponents I know view the directive as having had the teeth negotiated out of it while Chris Castle has opined that, for instance, YouTube will actually comply with the Directive shortly before Hell freezes over … 

One goal of the Directive was to have very large platforms, like YouTube, pay license fees to the copyright owners whose works contribute substantially to their revenue models. Thus, Article 13 (now 17) requires these large sites to develop technical measures to prevent and mitigate infringing uses, and this has been interpreted by many to mean that these platforms will implement filtering algorithms to identify protected works at the moment when a user initially uploads a file.*

For copyright owners, the proposal represents an important reversal of the current paradigm in which their works are chronically used without license, and they try to respond to every infringing use, one-by-one (a.k.a. the whack-a-mole problem), which is virtually impossible for a small, independent rightsholder. For the critics of the Directive, the assumed filtering at the point of user upload has been characterized as censorship because this would inevitably stifle all manner of legal uses before they are even published. One way or another, it should to be obvious why the status quo has been beneficial to YouTube and why the company used its platform to promote the censorship narrative against the legislation. 

In that regard I tend to agree with friends and colleagues who view passage of the Copyright Directive as at least a symbolic win for the democratic process itself. Certainly it is a refreshing change that the European Parliament was not bullied by Silicon Valley and was able to cut through a lot of the standard fear-mongering noise that was so effective against the SOPA/PIPA bills in the United States in 2012. Perhaps screaming #SaveYourInternet! is an exaggeration that has finally reached its limit.

But because I think the manner in which policy is debated is actually rather important, that brings us to the title of this post and the fact that it is apparently necessary to clarify that the 17th-century, English poet John Milton never said or wrote a bloody word about the internet. That may seem obvious to most, but not so much to history professor and novelist Ada Palmer, writing for the venerable Boing Boing.  In her post of March 24, she alleged that Article 13 of the EU Directive is comparable to European censorship of books in its ancient past.  She begins …

The EU Copyright Directive rekindles the oldest fight in the history of free speech debates, first waged by John Milton in 1644.  Then, like now, policy-makers were considering a radical change in censorship law, a switch from censoring material after it was published to requiring a censor’s permission to publish in the first place.

Nobody with the words professor and history next to her name should be allowed to write anything that misleading without rebuttal. And not wanting to bury the lede here, the most rudimentary response to articles like Palmer’s is that we cannot responsibly compare any proposal to require licensing of works by a for-profit corporation to a state’s authority to allow or disallow publication of works in the first place. There is simply no ideological or practical relationship between those two paradigms, let alone any relationship that spans from Milton’s role in the early phases of the English Civil War of the 1640s to the need to address various effects of internet companies in the present.

I’ll spare readers an attempt at historic summary, but England’s narrative vis-a-vis publishing, censorship, religion, and politics was—especially from the time Henry VIII broke with Rome—a schematic drawn by a psychopath. To give you some idea of what I mean, the fundamental language of the printing laws, beginning no later than 1557, hardly changed from iteration to iteration while the religious, social, and political postures of the state vacillated wildly for nearly 200 years.  

Hence, the 1643 proposal to end “great late abuses and frequent disorders in printing many false, forged, scandalous, seditious, libelous, and unlicensed papers, pamphlets, and books to the great defamation of religion and government” was not new language, but it had very different implications in Cromwell’s Parliament than it did under each of the precedent reigns of Mary I, Elizabeth I, James I, and Charles I before he was beheaded in 1649. 

So, Palmer is simply incorrect to call Parliament’s Printing Act proposal of 1643 a “radical change in censorship law.” To the contrary, it was a continuation of the same censorship laws being adopted under new management; and this was central to Milton’s complaint when he published his treatise Aeropagitica in 1644.  As both a literal and literary soldier for the cause of the Commonwealth, Milton saw the renewal of state publishing licenses as a needless reiteration of the royal prerogative to censor books (i.e. anathema to the anti-royalist cause itself). Notably, it was John Locke, father of the labor theory grounding one pillar of intellectual property, who inveighed against yet another renewal of almost the same law in 1694.

While it is true that Milton’s Aeropagitica is widely recognized as one of the most beautifully articulated defenses of free speech in our political heritage, it must remain in a context that is a very great distance from our modern conception of speech. For instance, Milton advocated allowing heterodox books et al to be published so that virtuous men would recognize the difference between good and evil; and while this is solid Puritan reasoning, it is not the underpinning of speech rights today—to say nothing of how far removed, say, “Paradise Lost” is from the average YouTube video.

Palmer makes a mistake common among academic critics when she draws straight lines between ancient Europe (usually England) and modern copyright. In fact, these lines are not only serpentine but highly fragmented. It would take a very long post to unpack all the moving parts, but just as Milton’s Aeropagitica should be seen as one ingredient in a philosophical soup that is many years away from tasting like the Bill of Rights, that’s more or less how we must view the practices of the English book publishers (the Stationers Company) and their relationship to the volatile English state between roughly 1557 and 1774.  

Simply put, only the state had the authority to grant (or not) permission to publish a work, while the Stationers’ monopoly on publishing was akin to the monopoly every other London guild—there were more than a hundred—enjoyed over its trade.  While there was intermittent correspondence between the publishing monopoly and state censorship, that narrative is a crazy quilt of inconsistency and has almost nothing to do with modern copyright law, which vests ownership in the expressions of individuals.

Yet, by invoking Milton, and even sprinkling the word Inquisition throughout her article, Palmer would have readers believe there is a through-line in the narrative—a kinship between copyright enforcement among private entities in the 21st century and government censorship of publishing in the 17th. There is not. And frankly, I am not sure which is a greater threat to rational public discourse—the hashtag that distills a complex issue into a lame bullet point, or the erudite article that sounds reasonable but is fatally inaccurate.  For sure, the democratization of the internet has given both more room to roam than either deserves, which does beg the question as to what this fight is really about.

* This is not mandated by the new statute and may in fact be how the platforms respond, but that’s a topic for a different post.

Update: Cy Pres & SCOTUS

Last October I reported that the Supreme Court granted cert in a case that might have led to a decision regarding complaints that cy pres awards in class action lawsuits wind up funding organizations anathema to the interests of the class. Specifically, I addressed the amicus brief filed on behalf of copyright owners outlining the ways in which, for instance, Google can settle a suit then fund various organizations (EFF, Public Knowledge, et al) that work to undermine copyright law.

Unfortunately, the Supreme Court vacated and remanded the case of Frank v. Gaos back down to the lower courts on a question of standing. So, there will be no decision in the near future on the broader question as to whether cy pres awards too often harm the interests of litigating classes. That said, Justice Thomas, in his dissent, wrote the following …

Whatever role cy pres may permissibly play in disposing of unclaimed or undistributable class funds (citation removed) cy pres payments are not a form of relief to the absent class members and should not be treated as such (including whencalculating attorney’s fees). And the settlement agreement here provided no other form of meaningful relief to the class. … the fact that class counsel and the named plaintiffs were willing to settle the class claims without obtaining any relief for the class—while securing significant benefits for themselves—strongly suggests that the interests of the class were not adequately represented. 

In short, because the class members here received no settlement fund, no meaningful injunctive relief, and no other benefit whatsoever in exchange for the settlement of their claims, I would hold that the class action should not have been certified, and the settlement should not have been approved. 

So, take that for what it’s worth if and when this subject returns to the Supreme Court.