— A weekly roundup of healthcare investigations
by Rachael Robertson, Enterprise & Investigative Writer, MedPage Today
August 12, 2026
• 3 min read
Stanford researchers successfully had an artificial intelligence (AI) model create a whole family of simple viruses that can only infect bacteria. While these particular viruses pose no threat to people, there have long been concerns about biosecurity risks tied to AI. Indeed, after OpenAI boosted the brain power of its chatbot last summer, hundreds of people around the world asked it how to make and deploy biological weapons and poisons -- and the models answered with step-by-step instructions. (Wall Street Journal)
