Google’s AI explains hallucinations, then names Anthropic as the culprit

conitzer · x · 2026-07-29

A screenshot shows Google’s AI answering a question about false statements by saying AI can hallucinate and then naming Anthropic as the company responsible for the system’s behavior.

The linked Substack post is a short “funny AI fails” note highlighting how the example accidentally illustrates the very concept it is trying to explain.

Original post →

More from Fun

Fun channel →