Words, words, words |
Strong resistance to AI among writers is understandable. But it obscures what we share with the machines: language itself
by Martin Puchner BIO
Studio Schwitters (2010), a sound installation by Pavel Büchler, plays the Dadaist sound poem Ursonate (1922-32) by Kurt Schwitters. Photo by Christophe Gateau/dpa/Getty Images
holds the Byron and Anita Wien Chair in drama and in English and comparative literature at Harvard University. As general editor of the Norton Anthology of World Literature, he has brought 4,000 years of literature to students across the globe. His books include The Written World: The Power of Stories to Shape People, History, and Civilization (2017), The Language of Thieves: My Family’s Obsession with a Secret Code the Nazis Tried to Eliminate (2020) and Culture: The Story of Us, from Cave Art to K-Pop (2023). He writes the Substack In Practice.
Since artificial intelligence went mainstream a few years ago, it has done double duty as a political personality test: tell me what you think about AI, and I’ll tell you who you are. Those worried about climate change focus on energy consumption. Those who denounce late capitalism see it as the ultimate example of corporate monopoly. Those concerned about racism have warned about AI biases. Those studying the effects of colonialism see it as yet another form of exploitation. And those tending toward doom have seen ChatGPT, Claude, Gemini and Grok as the four riders of the apocalypse.
People in the arts and culture have felt particularly threatened by AI because the technology seems to be coming for the things they cherish the most: the creative use of images, words, and ideas. The latter two, words and ideas, have been in the centre of the storm because generative AI is based on language and because ideas are closely associated with the words in which they are expressed. In response, writers have largely opted for resistance, defending the genuine creativity of humans against the machines. My social media feeds have been flooded with AI-slop gleefully produced and circulated by colleagues hoping to prove that AI can’t be creative. Let’s call this the Creative Resistance.
Based on my experience teaching and lecturing across the world, the Creative Resistance is strongest in North America, much less dominant in India, and still less in China and Korea, with Europe somewhere in between. When I taught a class on AI and creativity in Seoul last summer, with students from across Asia as well as Latin America, they had a single concern: please teach us how to use these tools effectively. The only person calling for creative resistance was an American student who had strayed into the class. In India, too, I’ve had many interactions with people in the arts that showed a less defensive, more exploratory attitude.
These anecdotal impressions are backed up by research measuring attitudes toward AI in different locations more generally. It’s tempting to speculate about reasons for this unequal distribution. The filmmaker Shekhar Kapur told me in a conversation that it was a matter of defending entrenched privileges: people in the Global South have less to lose and are therefore more open to new technologies. There might also be deeper philosophical and cultural attitudes, with Buddhist-inflected cultures less invested in the distinction between humans and non-humans.
No matter where it exists, the Creative Resistance is understandable as a reaction to a disruptive technology, but ultimately it gets in the way of understanding AI. Over the past three years, I’ve experimented with AI and have come away a cautious optimist. AI may well be terrible news for software engineers, but I think it’s an intriguing development for people who care about language and ideas – precisely the people who currently reject it the most.
Join more than 270,000 newsletter subscribers
Join more than 270,000 newsletter subscribers
Our content is 100 per cent free and you can unsubscribe anytime.
Our content is 100 per cent free and you can unsubscribe anytime.
The Creative Resistance has a point in that AI raises questions about what machines can and cannot do – but it also raises similar questions about humans. Many of the faults the Creative Resistance sees in AI – that it is predictable, that it merely recombines what is already there – are also true of many humans. In fact, much work in the arts of the past decades has questioned claims about human creativity. These include the cult of the lone genius; the role of institutions in shaping art; the reliance on tools and media; and collective modes of creative production. It’s been strange to observe how quickly writers and artists, when faced with the threat of AI, have forgotten all that and are now seeking refuge in very traditional views of human creativity. It’s almost as if we were back in the world C P Snow described in the 1950s, divided between avant-garde science and tradition-bound arts and humanities.
The main problem with the Creative Resistance is that it blinds us to what is most interesting about AI, namely, that it is based on language – something we share with AI. To be sure, machines and humans learn language differently, process it differently, and interact with it differently. But these differences don’t mean that only we humans really use language while machines just mimic it. AI agents are astonishingly effective users of language, though with different strengths and weaknesses from us. Acknowledging this doesn’t commit us to saying that AI is conscious, a famously difficult concept to define, or in other ways human-like. Indeed, one of the strengths of AI is that it uses language differently from us. But the point here is that language is the terrain on which humans and machines meet. Let’s call this the Shared Language Model, which also explains why humans can interact with AI in the first place.
Living with the Shared Language Model is something that we need to learn. Until recently, it was reasonable to assume that all speaking agents were humans (apart from trained parrots), hence our ingrained habit of attributing personhood to chatbots, with all the much-publicised consequences. But now we need to get used to having another language-user among us.
Many readers feel cheated when they learn that a text was written by AI. But will this attitude persist?
Many disciplines will be needed to........