Anthropic said the programming chatbot was exploited to help carry out attacks that “potentially” hit “at least 17 distinct organizations. OpenAI in June ...