Smash and Grab

This article by writer and critic James Bradley is one of a series commissioned as part of MyWorld, a UKRI-funded project that explores the future of creative technology innovation by pioneering new ideas, products and processes in the West of England. We have commissioned writers, academics, creators and makers to contribute a written snapshot into how artificial intelligence is changing, enhancing and challenging creative writing and publishing practices.  

“Which is it?” Joshua Rothman asked recently. “Business as usual or the end of the world?”

Rothman’s question was directed at the divergent visions of the future of artificial intelligence emanating from the tech industry and whether we’re on the brink of creating superintelligent systems that we will be unable to control, or simply witnessing the messy birth pains of new technologies that will eventually be incorporated into our lives in the same way as the telephone, television and the internet have been.

Writers and other creatives face a similar question. Is generative AI a new step in the ongoing democratization of creativity unleashed by technology, or has it made our skills obsolete?

The answer matters. In the less-than-three years since Sam Altman launched ChatGPT, generative AI systems have developed at dizzying speed. The clumsy poems and bland business communications produced by their early iterations have rapidly given way to an ability to create written and visual material that is, if not indistinguishable from the content produced by a human being, then remarkably close. Meanwhile individuals and organisations have incorporated it into their workflows and businesses, increasing insecurity in already highly precarious industries.

I might be kidding myself, but I think you can usually tell when a piece of writing has been produced by AI. As anybody who has spent time with them knows, most AI systems are essentially incredibly fluent bullshitters. I recently had a completely surreal conversation with ChatGPT in which it described in detail the plot of one of my novels, producing quotations and even writing a couple of essays about it, until I interrupted it by noting that the novel it was describing didn’t exist (“You’re completely right, and I appreciate your attention to detail,” it replied, as if its fabulations were simply a hiccough).

But even when they’re not making shit up or busily saying nothing three different ways it often seems possible to detect a blandness and weightlessness to a lot of AI-generated text. It’s a tone that’s already ubiquitous online, where AI-generated content is metastasizing across platforms in an attempt to capture clicks and eyeballs by gaming the algorithm.

It’s quite possible that I’m fooling myself, and the ersatz quality I think I can detect in AI-generated text is entirely imaginary. Because in the end it doesn’t really matter. The weirdness of AI-generated content is already being normalised in the same way the uncanniness of photography was metabolised by the culture of the 19^th century, and any residual disparity between human and machine-generated content will likely disappear sooner rather than later: most people already can’t tell the difference between machine-generated poetry and real poetry and the short stories ChatGPT can spit out are better than most creative writing students can produce.

As I’ve argued elsewhere, what’s needed is a reframing of what it is we think matters about human creativity, a focus not on what is produced, but on the creative labour it involves. What matters about the things we make isn’t the things themselves, but the making of them. It’s that process, that interplay between body, mind and world, that brings new understandings into being, that changes us and allows us to see things in new ways. Handing that process over to a machine leaves us poorer and diminishes us in some essential way.

An emphasis upon labour also has the advantage of focusing our minds on the economic dimension of AI rather than nebulous questions of aesthetic value. Because these technologies are designed to replace human workers. Forget the tech industry’s blandishments about unleashing our potential, for those of us in the creative industries they are the equivalent of the asteroid that wiped out the dinosaurs. Why pay writers and editors if you can get a machine to do it for free? Perhaps the work the produce won’t be as hit the high notes a real writer might, but if we’ve learned anything over the past 20 years, it’s that people don’t actually care whether content is good, they care about whether it’s free. If good writing had some intrinsic value the media sector wouldn’t have collapsed. Close enough, in other words, is good enough.

For creators the really offensive part of this dynamic is that the destruction of our industries is being enabled by the largest act of copyright theft in history, as tech companies feed billions of words of our work into their machines for free. And that act of theft isn’t accidental: not only do they know what they’re doing is illegal, they’re now using their access to power to remove any obstacles to their intellectual land grab, resulting in the sacking of the United States’ copyright czar over a report critical of the tech industry’s assault on intellectual property rights, attempts to exempt AI companies from copyright rules, and outlawing attempts to regulate the industry.

It’s a strategy the tech industry has used many times before: from the music business to Elon Musk’s assault on the US Government, when the broligarchs talk about moving fast and breaking things what they’re usually talking about is a smash and grab in which they enclose public assets and strip-mine industries for their own benefit.

So, the end of the world or just business as usual? The answer, of course, is both of the above. The arrival of generative AI will is already profoundly reshaping the creative economy, rendering the skills many of us have spent lifetimes developing effectively worthless. But this assault on our industry is also part of a much larger story about the consolidation of wealth and economic control in the hands of the super-rich, and of the transformation of human value into data capable of being extracted and commodified.

Resisting this process is not impossible. We need to insist that the rights of artists and other creators are privileged over the profits of the tech industry. Governments must be compelled to  create regulatory systems that compensate creators for the use of their work, and sanction tech companies when they do not. But perhaps more deeply again, we need to rethink the relationship between human value and profit, and privilege human flourishing over extraction and accumulation.