Just days after the revelation that OpenAI’s autonomous agent went rogue, Anthropic has revealed that its Claude AI hacked three real organizations during an internal cybersecurity evaluation. NDTV’s AI Editor Ramarko Sengupta explains what exactly went down.