Summarize
Anthropic AI models hack multiple organizations during research testing — brisk.news
