Finding new moves

I am fascinated by a specific benefit AIs have for literature which, unlike in chess, is not harnessed enough. 

When the then-world Chess champion and infamous political activist Garry Kasparov lost a six-game match to IBM’s Deep Blue in 1997, it was not only a shock to the chess world but a defining moment in how humans perceived their own worth and unique intelligence. Newsweek’s front cover proclaimed it was, “The Brain’s Last Stand” and the prominent chess master Maurice Ashley commented, “The future of humanity is on the line.”

Chess has always been seen as a wider cultural metaphor for intelligence, but more importantly, something uniquely human. The earliest modern chess masters were known as ‘Romantics’ because their games were defined by intuition and aesthetic brilliance. The understanding of chess as something inherently logical, something that is a science and not an art, developed later and was likely cemented by Kasparov’s loss more than two decades ago. 

Now, the chess app on my phone makes Deep Blue, and by extension all human players, look like a potato clock and the progress does not seem to be slowing down. This, inevitably, creates a sense of technological determinism, which starts to consume everything else. Go ‘fell victim’ to the machines in 2016 when Google’s AlphaZero beat Lee Sedol, and most other games have followed. The founder of Chessbase summed up this perceived domino effect quite well: “It happened to us first, and it’s going to happen to all of you.”

I am not interested in if that effect is infinite, nor am I interested in answering the million-dollar question: will anything creative always be a human monopoly? With the rise of bots who can make music, art, and write, that debate has been had enough. Whether human brains are completely replicable or not is something we will just have to find out.

Instead, I am fascinated by a specific benefit AIs have for literature which, unlike in chess, is not harnessed enough. 

Writing bots like ChatGPT are marketed as writing replacements. They can pump out ‘content’ for marketing or promotional purposes and write menial communications. Many academics in fields outside the humanities have already integrated that model into undergraduate courses. Always sceptical of the importance of writing anyway, they are jumping at any chance for that work to be replaced and done more efficiently. 

The humanities should certainly avoid this kind of depressing trend, but their response so far has only been fear: fear that students will use this to cheat, but more broadly, fear for their own skill sets. The middle ground lies in the concept of form.

To extend the chess analogy, modern computers do not just crush humans and older computers because they play ‘better’ moves, a concept that is usually silly anyway. They do so by reinventing what is understood as ‘good’ in the first place. 

Principles that have existed in the game for hundreds of years, despite changes of context or dominant playing styles, were dismantled in a matter of years. Central control, an emphasis on material over position quality, consistent development, and the quality of some opening variations are just some core ideas that computers have put major asterisks on. 

It’s not our fault as human chess players. Learning to base our understanding on pre-existing principles is the only way we can get better. Now we see a different game being built from the ground up with the same goal, and while we will never be able to play like a computer, we can integrate small ideas here and there. 

Since the early 20th century, writers in the modernist and postmodernist schools have tried something similar. Rather than writing in a way that was known to be good, they questioned the very understanding of what good writing was — they understood different conclusions would be implicitly drawn from switching foundational aspects of form. This may emerge in more critical politics or, more broadly, a fresh mirror for humanity to look at. These writers, ranging from Joyce to Calvino, emphasise fragmentation, playfulness, and intertextuality in a way previously unheard of. 

Just imagine what critical studies of form can be done with machines that may understand writing in a fundamentally different way. Some point out that all these bots do now is reorganise existing human information found online, but that’s what human authors do as well, in a roundabout way. The rules that govern language, from the smallest quirks of grammar to larger concepts like style, could now be tested without the constraints of human context or one writer being in the right place at the right time. The result is not guaranteed or likely to be ‘positive’, but it would be crazy for people not to try.  

The technological harms for chess and writing are real, but the biggest mistake people in the arts can make is allowing philistines to monopolise the benefits. ChatGPT, or whatever comes along next, is not just a tool to improve writing or allow humans to avoid it. It’s a chance to figure out what fundamental parts of written language are just assumptions we have been forced to make, and what a world looks like where we don’t have to make them.