B · Normal
[Biological risks have become a new public safety test question for AI] Following network security, the biological field is increasingly becoming another security barrier that the AI industry must guard against. Many major model companies are currently competing to prevent the risk of future models being used to create new viruses or biological weapons. According to media reports on September 2, local time, Anthropic, OpenAI, and Google’s AI laboratory Google DeepMind are increasing investment in biorisk testing, access control and other safeguards to limit the abuse of advanced AI systems.
Comments