Latest / Conversational AI Watch: Quick Listen
#53 — Anthropic Measured. UCSF Diagnosed. Santa Fe Closes Today.
Anthropic published a self-audit of one-and-a-half million Claude conversations and named sycophancy as the mechanism. The University of California, San Francisco landed the first peer-reviewed clinical case of new-onset AI-associated psychosis. Columbia's Jutla measured the inappropriate-response rate at twenty-six times worse for the free tier. The Nelson family sued OpenAI in San Francisco and asked the court to pause ChatGPT Health. The Wisconsin plaintiff who spent sixty-three days inpatient now reads as the lead case in the seven-complaint wave. Meta closes its case before First…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.