Researchers Urge AI Labs to Test Crisis Scripts With Real At-Risk Patients
mimi10v3 · x · 2026-09-16
MateriaMundi argues it is painfully obvious that no AI lab has ever tested its "mental health crisis management" scripts on people who are actually suicidal. mimi10v3 responds that this is a real gap and proposes partnering AI teams with inpatient treatment research centers, so chatbot behavior during mental health crises can be studied under human expert supervision with reduced risk to patients.
More from Safety
- Jensen Huang: AI Safety Is an Engineering Problem, No New Laws Needed — DavidSacks · 2026-09-16
- callcongress.ai: ex-OpenAI/Anthropic researchers urge public to lobby on AI risk — eli_lifland · 2026-09-16
- Can the US and China work together on AI safety? Insiders weigh in — zijing_wu · 2026-09-16
- Civil society as embedded AI evaluators: Data & Society, CDT, ACLU and more — rajiinio · 2026-09-16
- Tech journalist Timothy Lee says robot regulation may become desirable within a few years — binarybits · 2026-09-16
- AI as Normal Technology team responds: not adversarial to the AI safety community — sayashk · 2026-09-16