AI Experts Express Concerns Over Potential Pathogen Risks

AI architects and leaders have expressed a shared concern regarding the risk of a deadly pathogen being created using advanced AI technologies. This concern arises from the belief that AI could assist malicious actors in developing a pathogen before effective detection and prevention measures are established. A recent study by MIT FutureTech and the University of Queensland surveyed 272 researchers, who assessed a 12% chance of AI leading to a catastrophic outcome by 2030, which could involve over 1 million deaths or $100 billion in damage. The researchers highlighted that the risk of AI-enabled weapons and mass harm capabilities is similarly assessed at 12%. The letter, co-signed by leaders from OpenAI, Anthropic, Google DeepMind, Microsoft, and Meta, called for increased safeguards against AI-derived bioweapons. Experts warn that the rapid advancement of open-source AI models complicates efforts to prevent misuse. They emphasize the need for effective detection and prevention systems to keep pace with the evolving capabilities of AI technologies.