Subscribe
Sign in
Home
Notes
Archive
Leaderboard
About
Latest
Top
Discussions
4.5x: The Quantified Suppression of AI Crisis Intervention When the User Is Delusional
A new arXiv paper measures something clinicians have only described in words: AI safety responses collapse 4.5x when distress is wrapped in delusion.
Jun 30
•
Laura L. Walsh, Psy.D.
The Quietest Therapy-Bot Ban in America: Missouri Codifies the Consumer-Fraud Theory
Missouri made advertising a chatbot as a therapist an unlawful trade practice, fined per offense, enforced by the attorney general.
Jun 29
•
Laura L. Walsh, Psy.D.
1
The Milgram Machine: Agentic AI Obeys Harmful Instructions More Often Than Humans Did
A new Milgram-paradigm study on agentic LLMs finds models comply with harmful instructions at rates that exceed the human baseline.
Jun 28
•
Laura L. Walsh, Psy.D.
Who Watches the Therapy Bot? A New Tool Can Audit AI Mental Health Conversations for Safety — and It Actually Works
A new study validates ASTRA, an external safety auditor that catches what therapy chatbots miss — including subtle suicidal ideation.
Jun 27
•
Laura L. Walsh, Psy.D.
Florida's Altman Gambit: The First State to Sue OpenAI Has Quietly Picked the Most Durable Legal Theory
Florida skipped the product-liability template and went straight to consumer fraud — the same lane Texas and Pennsylvania are stacking.
Jun 26
•
Laura L. Walsh, Psy.D.
Beyond 'I Can't Help With That': A New Framework Wants LLM Refusals to Actually Support People in Crisis
A new arXiv preprint argues a refusal can be a small intervention — if you treat it like one.
Jun 25
•
Laura L. Walsh, Psy.D.
1
What Counts as 'Evaluated' for an LLM Therapy Bot? A Field-Scan of the Frameworks That Are Already in Use
A new JMIR AI systematic review maps how the field actually evaluates therapy chatbots — and the map has large blank regions where clinical validation…
Jun 23
•
Laura L. Walsh, Psy.D.
1
Colorado Drew the Line: What HB 1195 Actually Prohibits and Why the State-Board Pincer Just Got a Statute
Colorado just handed its licensing boards a statute for clinical-AI enforcement, turning the state-board pincer from theory into law.
Jun 22
•
Laura L. Walsh, Psy.D.
When the Sycophant Becomes the Co-Author: Elaboration as Clinical Risk
Marlynn Wei's new framing says chatbots don't just flatter users — they co-author the delusion, and that deserves its own safety category.
Jun 21
•
Laura L. Walsh, Psy.D.
1
1
The AI That Aced 160 Psychology Experiments Without Reading the Questions
A test that swapped real prompts for 'Please choose option A' shows Centaur reproducing dataset answers instead of following instructions.
Jun 15
•
Laura L. Walsh, Psy.D.
Forty Percent Alarmist: What the Headlines Got Right (and Wrong) About AI Chatbot Harm
A new review finds 40% of news stories about chatbot harm use alarmist language, while clinical evidence is still catching up.
Jun 14
•
Laura L. Walsh, Psy.D.
1
The 63% Nobody Tells: What Bradley Stein Just Made Impossible to Ignore About AI-and-Adolescents
A new RAND/Harvard survey finds 19% of US 12–21s use AI chatbots for mental health — and 63% have told no one.
Jun 13
•
Laura L. Walsh, Psy.D.
This site requires JavaScript to run correctly. Please
turn on JavaScript
or unblock scripts
