The controversy that revealed how AI has become a red rag to a bull for many authors

American novelist Curtis Sittenfeld recently ran an experiment in The New York Times where she agreed to write a short story – “an ideal beach read” – in competition with an AI bot given the same brief. Readers gave prompts, including “lust”, “regret”, “kissing”, “middle age” and “flip-flops”. Sittenfeld took a few weeks to write her story; the bot took a few seconds.

It was good to see such a light-hearted experiment take place in a world where AI is getting a darker and darker reputation with writers. Witness the kerfuffle over National Novel Writing Month, better known as NaNoWriMo. Since 1999, this nonprofit organisation has challenged participants to write 50,000 words of a novel in November. Hundreds of thousands have taken part and many have gone on to find publication, or at the very least to produce a story they can share with family and friends.

The use of AI is causing controversy in the publishing industry and beyond.

But now NaNoWriMo is reeling from a social-media meltdown. Writers on its board have resigned, a sponsor has backed out and arguments are still raging. All because it put out a statement aiming to be neutral about participants using AI, and saying that condemning the technology would amount to ignoring “classist and ableist issues”.

Clearly any mention of AI is a red rag to a bull for many authors. One of the first to resign from the board was bestselling YA novelist Maureen Johnson, who’d supported NaNoWriMo since its early days. It encouraged community and creativity, she said to The New York Times : “What I saw in their statement was the opposite of that.”

Another board member who resigned, fantasy and YA writer Daniel Jose Older, complained about ProWritingAid, an AI platform sponsoring NaNoWriMo this year. He said the organisation had taken “a wild and ridiculous stand in favour of generative AI” and its position was “vile, craven and unconscionable”. Then a sponsor, the writing software company Ellipsus, withdrew its support.

In full damage control mode, NaNoWriMo put out another statement. “We apologise that our original message was unclear and seemingly random,” it said in part. “Our note on ableism and classism was rooted in the desire to point out that, for people in certain circumstances, some forms of AI can be life-changing.” But this second post did little to halt the pile-on. Even disabled writers took offence: one, Laura Elliott, wrote on X: “Encouraging AI is a slap in the face to all writers and this excuse is appallingly ableist.”

Writers have already made good use of AI as a research tool, but they also have good reason to suspect other AI uses. Famous authors including George R. R. Martin, Jodi Picoult and Jonathan Franzen sued OpenAI (the developer of ChatGPT) last year for allegedly using their work to train its artificial intelligence tools, and some countries, including Australia, are alarmed at how such training could infringe authors’ copyrights.

Meanwhile, what of Curtis Sittenfeld’s experiment? I read the first few paragraphs of both stories and it was immediately obvious who, or what, had written each one.

As Sittenfeld says, individual sentences in the bot’s story are fine. But if you take the story cumulatively, “there’s not any real emotion or anything that makes it feel very specific… I feel like there is some sweatiness or nervousness or yearning that I think is actually inside of me and is in my fiction and is in a lot of people”.

The soul of the writer, perhaps? We will have to wait to see if AI can develop to the point where it can reproduce that quality. Maybe it never will. Meanwhile, writers are probably right to beware.

[e-mail]

A cultural guide to going out and loving your city. Sign up to our Culture Fix newsletter here.