A fresh trial-court ruling in Massachusetts sheds light on rippling effects from hallucinated support in filed briefings. In Wilder v. Harvard College, an attorney seeking admission pro hac vice in Massachusetts was denied in part due to hallucination sanctions in another case in another jurisdiction. Today’s issue includes Supplemental No. 02: Wilder v. Harvard College, an analysis of how AI sanctions may follow a lawyer.

Medical and science outlets are starting to cover the emergence of “AI therapy,” including chatbots made for mental-health support. In Issue 9, we noted Pennsylvania’s lawsuit against a company whose chatbot allegedly presented itself as a psychiatrist and supplied an invalid license number. Now, recent studies suggest purpose-built AI chatbots can reduce some anxiety and depression symptoms (JAMA Network Open; BMC Psychiatry), while performing better at handling explicit suicide prompts than subtle or slow-building risk cues.

Billions of dollars could go to U.S. intelligence agencies to fund the buildout of secure AI. (GV Wire)

One of the nation's elite law schools banned AI for most graded academic work. (ABA Journal / Source Link)

A major U.S. AI company is offering rewards to researchers who break its safety rules. (OpenAI)

Last week marked the start of enforcement for new federal legislation targeting AI-generated sexual “deepfakes.” (FTC)

Local governments in Arizona and New York are being pressed over traffic-camera contracts. (KJZZ / WAMC)