●What dangerous things have AI models done? Anthropic says its AI models have been used for harmful activities, including cyberattacks and fraud. The company also found attempts to use AI for research that could help make biological weapons. Anth...