AI safety researcher lists misdiagnosed patients, phishing-hit firms as AI's casualties

sarahbmyers · x · 2026-09-13

In her thread criticizing AI safety standards, Sarah Myers names the concrete victims: patients misdiagnosed by AI, companies facing wave after wave of AI-generated phishing attacks, and civilians wrongfully targeted by AI-enabled weaponry. She argues these safety failures would be addressed if there were a tolerance for declaring a system has too high an error rate and too little substantiated benefit.

Related event: AI Safety Researcher Warns 'AI Exceptionalism' Undermines Product Safety Standards(2 posts)→

Original post →

More from Safety

Safety channel →