Anthropic anunță că a blocat folosirea modelul său de AI pentru dezvoltarea unei arme foarte periculoase

Anthropic has reported blocking attempts to misuse its AI model for developing dangerous biological weapons, revealing five case studies of potential misuse. The company identified 35 distinct research efforts that raised concerns, although it could not confirm malicious intent. The report highlights the risks of AI models being used for harmful purposes, emphasizing the need for protective measures. Anthropic noted that the same information could also lead to beneficial outcomes, such as vaccines. Specific cases included research on avian flu and gain-of-function studies.