Top Insights in AI ETHICS Just Add Humans: Oxford Study Underscores the Missing Link in Chatbot Testing A new Oxford study reveals that relying solely on chatbots for self-diagnosing medical conditions can lead to worse outcomes than traditional methods, emphasizing the critical need for human involvement in healthcare AI. The findings highlight that while AI tools are helpful, they cannot replace professional medical judgment.  RESEARCH LLM Combo Achieves Superhuman Performance in Medical Research A powerful AI combo—using GPT-4.1, o3-mini-high, and Gemini 2.0 Flash—is transforming systematic reviews by completing 12 years’ worth of Cochrane research in just two days with accuracy surpassing human experts. This breakthrough enables rapid, reproducible, and scalable medical evidence synthesis, promising faster and more reliable insights for healthcare.  FUTURE OF AI Terence Tao on How AI Could Solve the Toughest Math Problems In a deep conversation on Lex Fridman's podcast, legendary mathematician Terence Tao discusses the future of AI-assisted theorem proving, its potential to solve famous conjectures, and how artificial intelligence could help tackle the most complex problems in math and physics.  | 
@gdb 
CBS Mornings 