By Daniel Thilo,Jonas R. Kunst,Gary Marcus
Publication Date: 2026-01-23 10:20:00
A joint essay with Daniel Thilo Schroeder and Jonas R. Kunst, based on a new article about swarms with 22 authors (including me). has just been published in Science. (A preprint version is Hereand you can see WIRED’s coverage Here.)
Automated bots spreading disinformation have been a problem since the early days of social media, and malicious actors have quickly resorted to LLMs to automate the generation of disinformation. But as we argue in the new article in Science, we foresee something worse: Swarms of AI bots working together.
The unique danger of a swarm is that it acts less like a megaphone and more like a coordinated social organism. Previous botnets were simple-minded, mostly just copying and pasting messages on a large scale – and in well-studied cases (including the Russian IRA action on Twitter in 2016) their direct persuasive effect was difficult to detect. Today’s swarms now emerging can coordinate fleets of synthetic personas – sometimes with stubborn…
