Shekhar Natarajan

The AI That Chooses to Be Good

Shekhar Natarajan said Traditional AI safety methods often rely on training models using techniques such as reinforcement learning from human feedback or external safety rules. Mar 5, 2026
Apple Chat GPT

'Deep respect for safety': Sam Altman on OpenAI's classified Pentagon deployment

The move comes amid reported disagreements between the Pentagon and Anthropic over limits on military applications of AI. OpenAI said it will implement technical safeguards, deploy field engineers, and ensure the models operate only on secure cloud networks, while maintaining that its safety commitments remain intact. Feb 28, 2026