Technology
IndiaAI COO seeks Indian seat in AI safety talks
Business Standard··5d ago
IndiaAI Mission COO Sudeep Shrivastava said AI safety rules should emerge from consultations that include India, not only the United States and China. He spoke at India Mobile Congress 2026 in New Delhi. He said 13 projects already run under the Safe and Trusted AI pillar. IIT Jodhpur's Mayank Vatsa said India still lacks indigenous AI safety benchmarks.
Prism
What It Means For You
- Indian firms may face AI safety rules written without New Delhi in the room.
- Thirteen Safe and Trusted AI projects are the current pipeline cited by the mission.
- Benchmark gaps flagged by Vatsa affect how Indian systems are tested.
What's Happening
- Shrivastava pointed to a recent US-China leadership meeting on AI safety channels.
- Those governments agreed to further dialogue, with the next round set for November.
- He also cited Anthropic's Mythos model and OpenAI agent attacks on Hugging Face.
Who Spoke and Where
- The remarks came at India Mobile Congress 2026 in New Delhi.
- Vatsa said no scientific models use benchmarks developed independently in India.
- Shrivastava said AI change is week-to-week, unlike Industrial Revolution pace.
all-news
