A critic who hates chatbots tried dating one. Here is what she found.
Writer Lauren Oyler went in sceptical and came out with something stranger than a verdict. Her essay, adapted for audio, is a rare honest look at what it actually feels like to talk to an AI.
Key points
- Writer Lauren Oyler published a personal essay about her experience using an AI chatbot, first in the Yale Review and later adapted as a Guardian audio piece.
- Oyler began the experiment as a self-declared opponent of AI, describing herself as "repelled" by the topic.
- The piece explores whether genuine intellectual engagement with a large language model, the technology behind chatbots, is possible or meaningful.
- The audio version was read by Kate Handford and released by The Guardian as part of its long-read podcast series.
Lauren Oyler does not like chatbots. She said so plainly, at the start.
Yet she spent enough time talking to one to write at length about it. The resulting essay, first in the Yale Review and later adapted by The Guardian into an audio piece, isn't a conversion story. It's something more uncomfortable than that.
Could an AI, a computer program trained on vast amounts of human text to hold a conversation, earn her genuine attention? The honest answer the essay seems to arrive at: sort of, sometimes, in a way that left her unsettled.
That feeling deserves serious attention. Most writing about AI chatbots lands in one of two camps: breathless enthusiasm or flat dismissal. Oyler sits somewhere harder to name. She notices when the machine says something that almost lands, and she notices the gap between almost and actually. For readers who've tried a chatbot once and felt vaguely odd afterwards, that gap is probably familiar.
What she doesn't do is pretend the technology is neutral. A large language model doesn't understand you. It predicts what words should come next, based on patterns in the enormous amount of text it was trained on, and can sound warm or even wounded. None of that is feeling. It's pattern-matching at a scale that mimics feeling well enough to be confusing.
That confusion is the real subject of the piece.
Should ordinary people worry about getting attached to AI chatbots?
Attachment is already happening. Millions of people use AI companions, chatbots designed specifically for emotional conversation, and some report genuine distress when those services change or shut down. Oyler's discomfort isn't a niche literary reaction. It points at something real.
If you or someone you know spends significant time talking to an AI for emotional support, it's worth asking what happens when the service disappears or the company pivots. Human support networks, whether friends and family or professional counsellors, don't vanish with a terms-of-service change.
Oyler's essay doesn't tell you what to do. It tells you what it felt like to pay close attention. That's its value: not a verdict, but a careful, slightly queasy record of what actually happens when a committed sceptic sits down with the machine and tries to be honest about what she finds.
