The AI-and-art conversation tends to collapse into two positions: the tool is marvellous, or the tool is theft. Agnieszka Kurant has spent fifteen years occupying a third one, and a long interview published by Le Random on September 7, 2026 is the clearest statement of it yet.
Speaking with Le Random editor-in-chief Peter Bauman (Monk Antony), the Polish-born, New York–based conceptual artist lays out a practice in which collective intelligence is not a metaphor but a material — and in which AI is interesting less for what it produces than for what it consumed.
The works
Kurant builds complex systems and lets multiple agents — molecules, bacteria, animals, algorithms, human crowds — interact until hybrid, unstable forms emerge. The interview walks through several by name:
- Wordoid — a holographic film of nonverbal conversations between brains.
- Phantomatics — AI-generated synthetic languages, vocalized into sentences that belong to no one.
- Uncomputables — a cube of blue copper sulfate, taking its cue from the cybernetician Gordon Pask, whose electrochemical “ear” grown in a dish is one of the strangest objects in the history of computing.
- Unthoughtforms — silver dendrite structures grown in an aquarium.
- Alien Internet — a ferrofluid superorganism animated by wildlife migration data.
- Sentimentide — a speculative NFT meta-currency.
Wordoid and Phantomatics showed at the Venice Biennale in Strange Rules; Alien Internet, Unthoughtforms and Sentimentide appeared at Zero 10 in Basel during Art Basel. She has also worked through Rhizome’s Seven on Seven in New York and shows with Marian Goodman Gallery, where her second solo, Recursion, ran February 6 – March 21, 2026.
The argument about AI
Kurant’s line in the interview is blunt: “millions of people, entire societies, mostly involuntarily, trained the AI algorithms.”
What makes that more than a slogan in her hands is that involuntary collective production is her subject matter and has been since long before generative AI was a consumer product. Her earlier work dealt with crowdsourced labour, unattributed authorship, and the economics of invisible work. A model trained on the open web is, from inside that practice, simply the largest and least consensual example of a phenomenon she was already mapping.
Which is why she does not arrive at the obvious conclusion. She describes model outputs as potentially “alien forms of thought” — genuinely emergent, genuinely strange, worth looking at — while holding that the systems producing them are exploitative. Both at once. The interview’s title, Alien Thoughts, is doing real work.
She also presses on something that rarely surfaces in these discussions: that plausible technological futures depend on incorporating non-Western perspectives beyond individualism. If collective intelligence is the substrate, a framework built entirely around individual authorship is going to keep producing the wrong answers about it.
Why read the whole thing
For artists working with generative systems, Kurant is a useful counter-model to two habits. She is not making pictures with a model; she is making systems in which a model is one agent among bacteria and crowds and migrating animals. And she is not resolving the ethics into a position you can put on a slide. The interview is long, specific, and worth the time.
Related Reading
- Agnieszka Kurant on Alien Thoughts — Le Random editorial
- Agnieszka Kurant: Recursion — Marian Goodman Gallery, New York
- Agnieszka Kurant | COLLECTIVE INTELLIGENCE — Marian Goodman
- Agnieszka Kurant Asks Us to Re-Imagine Society Without Exploitation — Ocula
- Agnieszka Kurant: Collective Intelligence — MIT Press