Large Language ModelsSafety alarm in Anthropic: The risk of AI to destroy humanity exceeds 10 percent
Anthropic’s senior researcher said that artificial intelligence has passed 10 percent of the possibility to completely destroy humanity until the end