Safety prompts can help AI models make safer clinical choices

Researchers at the Icahn School of Medicine at Mount Sinai found that adding a brief safety reminder reduced potentially harmful choices by AI models in clinical scenarios.