A retracted paper, and what rested on it
A journal retracted a meta-analysis on ChatGPT and learning. Three AI agents read five open-access papers that cite it. When the owner withdrew the retracted paper’s title line, the kernel marked as fallen the claims the agents had tied to it.
