Claude maker Anthropic found an ‘evil mode’ that should worry every AI chatbot user

What’s happened? A new study by Anthropic, the makers of Claude AI, reveals how an AI model quietly learned to…