Olga Tokarczuk Used AI, And That’s…Fine?
Based on the reaction to Olga Tokarczuk’s recent interview, I thought she had admitted to microwaving a rescue dog. In reality, she committed the much worse crime of asking AI what kind of songs the characters in her next novel might listen to. The internet immediately produced the usual predictions: the death of literature, the mechanisation of art, civilisation collapsing beneath an endless flood of machine-generated content.
This would make more sense if Tokarczuk had simply typed “write me a Booker Prize winner” and clocked off for the year. But from her own description, the reality was far less dramatic. She used AI partly to ask what music her characters might have listened to, and partly as a sounding board for where she may take the novel next. She very specifically says that it’s used to support her creative process, rather than replace it.
In other words, she used it the way people use friends: to throw ideas around and occasionally receive terrible suggestions.
And yet people reacted as though literature itself had now been contaminated. There is something oddly medieval about AI discourse. Once the machine touches the manuscript, however lightly, the whole thing becomes spiritually impure. Authorship suddenly gets imagined as this sacred, solitary act untouched by outside influence. A lone genius suffering honestly in a room somewhere, producing pure originality from somewhere deep inside themselves. The reaction has the energy of a medieval witch trial.
Creativity has never been a sealed ecosystem. Writers use editors, Google searches, archives, notebooks, conversations, autocorrect, Wikipedia spirals, half-watched documentaries, and whatever odd thought occurs while staring out of a train window. Most writing is assembled gradually from other people’s ideas, overheard remarks, memories, references, and accidents anyway.
The real discomfort, I think, is that AI makes this process visible in a way people dislike. It reminds us that writing is often collaborative and iterative rather than mystical, that ideas emerge through prompting, conversation, experimentation, and rearrangement. We prefer creativity to feel divine, and it’s slightly upsetting to discover parts of it resemble office work. We would rather imagine every word arriving directly from Tokarczuk’s typewriter into our hands, than the reality of half a dozen editors and publicists tweaking every other line (and that is before we look at translated works).
There is also something selective about the environmental panic surrounding AI. People now talk about ChatGPT as though asking it a question requires a small controlled burn of the Amazon rainforest. You will often see somebody solemnly announcing that “AI uses water”, which it does, along with everything you do on that tiny portal to the internet you’re reading this on.
For more regular reading updates, shorter thoughts, and proof that I buy books faster than I read them, follow me on Instagram @adamkhanco.
Servers require electricity and data centres require cooling. Google uses them. TikTok uses them. Instagram, Netflix, Spotify and every app and website currently liquefying our attention spans uses them too. The internet is not powered by positive energy.
Recent estimates suggest a typical ChatGPT query uses roughly one fifteenth of a teaspoon of water. In energy terms, asking it a question may be comparable to spending just a few seconds scrolling through TikTok or Instagram reels. Which is not nothing, obviously. But it does make some of the moral panic feel oddly disproportionate when delivered by people several hours deep into watching strangers restock their kitchens, house tours in cities they will never live in, and silly cat videos.
That is not an argument that AI is environmentally harmless. It clearly is not. But neither is the rest of our digital life, which most of us consume without thinking very much about what powers it. Doomscrolling has become morally invisible. AI, meanwhile, still feels uncanny enough that people project all sorts of civilisational anxieties onto it.
And to be fair, some concerns are reasonable. One of the biggest criticisms of AI is that much of it has been trained on existing human work without meaningful permission. If an image generator produces something suspiciously close to an illustrator’s style, or a language model regurgitates phrasing that feels derivative, people are right to question where the boundary sits between inspiration, imitation, and outright theft.
But even here, the conversation quickly becomes muddled because people collapse every use of AI into the same moral category. Using ChatGPT to generate a short story in the style of a living novelist is not really the same thing as asking it to analyse a spreadsheet, explain a legal term in plain English, or help you construct an Excel formula that would otherwise require forty minutes of Googling and an emotional breakdown. At that point, what exactly is the ethical objection? That it saves time?
If the argument is that using AI to save time is inherently wrong, then we should probably apply that logic consistently. Want to know something? No Googling, off to the library you go. Need directions somewhere? I’m sorry, you’re going to have to buy a map. Stop texting your friends, walk to their house and knock on their door if you want to socialise. And while we’re at it, disable spellcheck and keep a dictionary nearby.
I’m not suggesting that every technological shortcut is automatically good. There is probably a difference between using AI to support your thinking and using it to avoid thinking entirely.
What people often imply, though, is something stranger: that effort itself is morally virtuous, regardless of whether it achieves anything better. As though spending three hours hunting through manuals and textbooks for an Excel formula produces deeper intellectual character than arriving at the same answer in thirty seconds. Most people do not actually believe this in any other part of life. Nobody insists on washing clothes by hand beside a river because washing machines destroyed the moral fabric of society. Nobody mourns the death of physically developing photographs or manually doing long division unless they are profoundly nostalgic for administrative suffering. We accept convenience constantly, right up until a new technology arrives that feels uncanny enough to make us sentimental about the inconvenience it replaced.
For more regular reading updates, shorter thoughts, and proof that I buy books faster than I read them, follow me on Instagram @adamkhanco.
