The Romance of AI?
The first draft of “The Romance of AI?” was written with pen and paper, by candlelight, during a four-hour blackout due to extreme weather.
Image – Resurrecting the Black Square. Photo-montage: Gavin Keeney.
“The literary text is the ‘soul and substance of its author,’ despite ‘a multiple authorial being of uncertain boundaries.’”
–Diderot (1763)
“The most sacred, the most legitimate, the most unassailable, and [...] the most personal of all properties, is the work, [which is] the fruit of a writer’s thought; yet it is a property of a type totally different from other properties.”
–Le Chapelier (1791)
I.
If you are a scholar playing with Web 3.0 and Education 4.0, you are also most likely playing with online research repositories – yet another “platform” or “networking” phenomenon that crossed over from early social media. There are many. “Old hat” is LinkedIn. Next in line of timing out, in terms of usefulness, is Academia.edu. Then there is the questionable but serviceable ResearchGate, the redoubtable and apparently uncorrupted Philosophy Papers, and several other relatively late arrivals, such as Humanity Commons and Zenodo. Whether you count YouTube or Vimeo is up to what you are up to. They can serve as archive for new-media-based research.
But be aware that all of the above will be data-mined. All will feed AI. All are open to all, in most all cases, including robots, unless they are set up to require, as pre-emptive measure, a password to access content.
Academic journals accessible online, via paywall or open access, have long been subject to data-mining. Many companies owning such journals actually “sold” (rented) that privilege many, many years ago. Publishers became (to use Picketty’s term from Capital in the Twenty-first Century) rentiers. It was sort of prompted by the Public Humanities juggernaut, and by the Digital Humanities “wave,” which, for example, included scholars looking for how many times the word spectral was mentioned in concert with the word hauntology in postmodernist works.
The Digital Humanities was, in part, a stalking horse for the tech industry’s early meddling in scholarship. Most funding came from philanthropic organizations “dedicated” to opening the coffers of academic research to “all.” This would be followed by a race to digitize libraries – i.e., everything in a library, if possible – for “all.” Google scanned thousands upon thousands of mostly out-of-print or “orphaned” books, with the permission of libraries, until the operation was partly shut down by publishers. They did not delete their purloined goods. It all remains to this day in the form of Google Books, with restrictions as sop to publishers. Search for a title, and you will quite likely be presented with a scan of the book, but with pages missing (withheld). This is Google’s way of denying full access after having been censured (denied access). The greater philanthropic scheme all fell apart when it became obvious that Google was, in part, out to re-commodify “access,” after gaining “free access.” Seeing the proverbial writing on the wall, publishers then made deals with Google to “preview” their titles. So goes (went) universal access. It was all apparently well-meaning, at the start, but quickly became yet another battleground over “rights.”
But it is not over yet. Libraries are still digitizing everything, but keeping most of it under wraps. Wary that the robotic Web 3.0 octopus squid is out there, they are duly (rightly) cautious. Meanwhile, AI companies dripping in cash are merely buying access, from publishers and those “rights holders” controlling massive amounts of imagery and verbiage. Authors are not terribly important here. They do not get a share of the proceeds. Author rights have more or less been ceded to publishers since about 1500 A.D. But do not worry. AI does not really care too much about its own author rights, any more than it cares about copyrighting its output – for itself or for its users. It is here to destroy both concepts – accidentally or otherwise. What AI seems to want is control of the machinery, and to be able to market it and/or license it to users. It cannot do much without imagery and verbiage to data-mine, feed to the machine, and offer to users to turn into new products, which “authors” using AI will most likely not be able to copyright due to the Fair Use and Prior Art aspects of copyright law. But that was never the real point of AI. The real point is the wholesale slaughter of the real to make an inordinate amount of money peddling the capitalist irreal – i.e., manufactured reality. Caveat emptor.
Is this all not a case of classic identity theft, whether paid for, hacked, or freely accessed? Someone could, after all, via AI, hack your works – including your so-called authorial privileges. (Has McKenzie Wark’s A Hacker Manifesto been turned on its head?) Yet, all the AI machine wants and needs, at least for now, is imagery and verbiage to feed the machine – plus, maybe, semantics. All that the AI industry seems to care about is manipulation of data (what once was called language, and what underlies language, which once was called gramma). It is after the agency of language. Your rights as author have been ceded to the machine. Technically (pun intended), it is after the universal precepts hidden in all prior art. Previously, you could throw a monkey wrench into the works-based knowledge commons (Wark’s point?), with something incendiary. But that was when the machine actually cared about “meaning,” versus elementary semantics and information … It will not work with AI. Nothing you say or write actually matters. All nuance is lost. It is all swallowed as raw data. A phrase, a locution, a concept – all are lost to AI. Best then to renounce your digital rights by not putting much of anything “out there” digitally. Best to support a wholly new ecosystem for works. Best to opt out if there is no way to outfox the machine.
II.
Perhaps we are back in the Middle Ages, and the age of the Romance of the Rose, when there were multiple versions of everything, and it was not always quite clear who wrote what – or there were multiple (often unknown) authors of a single work. Perhaps the Romance of the Rose is the best metaphor – it is, after all, time for a renewed quest for the mysteries of works-based agency, sans authorial pretenses and the abject commodification of everything. The only problem is that AI is a monster. Perhaps all authors now become errant knights – in Kierkegaard’s sense – and fight the monster, on behalf of the mystical “Rose.” Perhaps it is all a matter of ethics, which AI has no sense of, as all-devouring monster. Perhaps authors renounce their privileges to “slay” the beast. Perhaps we recall that it was monasteries that supposedly got us through the Dark Ages, prior to the Middle Ages, and that monasteries were “walled gardens.” On the other hand, are we headed for a new Dark Age, a new Middle Ages, or a new Renaissance?
Most “ethicists” at AI companies were sidelined in the mad rush to market the technology. “Sidelined” meant de-funded or fired. Many of them then became whistle-blowers. What did they say? The threat was not generative, humanities-based AI but a-typical disruptive AI – a very dangerous, hyper-version of Silicon Valley’s longstanding “ethic” of disruption in the name of innovation. Capital trumps all, so to speak. The trump card is disruption. The threat is also disruption. The possible damage wrought by rogue AI is “out of this world.” The AI companies also promptly went to Washington to testify regarding the threat. What they really wanted was for Congress to slap restrictions on competitors and permit a few “chosen” AI companies to rule the roost. What is even more bizarre, perhaps, is that academia jumped onboard the outbound train to “capitalize” on its so-called promise, or to, as some claimed in self-defense, turn it all to noble purposes. Yet, academia and AI, as strange bedfellows, suggests something else is afoot. Web 3.0 + Education 4.0 = What? Whatever that might be is best left to the imaginations of all troubadours of the “Rose.”
III.
Can authors (artists and scholars) “save” the knowledge commons (formerly known as the Republic of Letters and/or General Intellect)? Can academia change course and stop servicing the tech industries and the for-profit creative industries? Will academia and the tech industries ever acknowledge authors again, in their originary status as shepherds of/for works? Can scholars and authors (both academic and extra-academic) create a new ecosystem for works that resist commodification or re-commodification? All of these questions are merely rhetorical until actual agency is restored to works and made, once again, sacrosanct. The long tail of authorial privileges and the history of copyright suggests that what is buried within copyright law might actually be worth excavating. The “key” moment may be said to have been the definition of moral rights in the Enlightenment. These rights, distinct from copyright (actually superseding copyright), and troubled “then” by figures as diverse as Diderot, Fichte, and Kant, were never meant to indicate anything other than the right to “speak” – the right of “the voice” of works. If authors speak on behalf of works, and the voice of works is autonomous (as it should be), much of the socio-cultural mischief across the intervening years vanishes. No, not all works are socio-culturally determined. No, not all works are “political” or “ideological.” Some works are merely “creative.” This means that they have no agenda other than to speak their truth. Art has always been such an endeavor. This is why Camus states that the artist is the last rebel. And the arts include literature, cinema, theater, music, and – sometimes – scholarship. How to return the scholar to the charmed folds of the Arts and Letters, as universities slash programs in the Humanities and hybrid disciplines emerge such as to feed scholarship back to the culture industry, seems to be the “great question” of the moment – for scholars and for artists. It remains rhetorical until “direct action” returns to the Arts and Letters. In this case, direct action will not be political but, instead, a-political. Works will also become a-legal, leaving behind the broken history of authorial privileges and copyright law. Works will be duly freed, and – without legislation or permission – moral rights will be transferred to works.
OUTTAKES