Cambridge researcher debunks misread OpenAI self-replicating prompt claims

Cambridge's David Krueger accused AI news accounts including TheZvi of misreporting the OpenAI self-replicating prompt incident, clarifying it was not found replicating in the wild and that such multi-agent prompt infection attacks were already documented in an October 2024 paper.

2026-09-29 ~ 2026-10-01 · 4 related posts

Full story(10 episodes)→