ISAIL Global Statement on "Pacing the Frontier"

Since a lot has been stated on "pacing" the frontier, the Secretariat of the Indian Society of Artificial Intelligence and Law and the Global Advisory Council co-led by Ayushi Agarwal, Bogdan Grigorescu and Deepanshu Singh have been observing the events and policy-technical discourse around the 2 frontier labs in the US suggesting to essentially prevent open source AI and pace the frontier.

Since their claims are clearly unproven, despite the fact that real AI risks still exist, it is clearly discernible for the Secretariat to make an official statement now.

1. We welcome some rare efforts by the Indian mainstream media for not becoming a victim of AI hype, and addressing real AI risks and the problem with valuations around these "frontier" AI labs, after a long hiatus.

2. We welcome the long statement of David Sacks (https://x.com/DavidSacks/status/2098973625252708460), who is co-chairing the United States President's Council of Advisors on Science and Technology (PCAST) on Dario Amodei's essay entitled "We Must Pace the Frontier" (https://darioamodei.com/post/we-must-pace-the-frontier). We don't expect Anthropic India to have a nuanced take on this either.

3. That being stated, technical and non-technical AI risks exist, and we therefore appreciate the interest of major US-based former regulatory officials, from Lina Khan to even David Sacks to espouse upon the necessity to promote more independent evaluators beyond METR. This is a suitable move and we agree with the AI Evaluator Forum and Miles Brundage's perspectives as well.

4. There are real-time AI agents-associative IP risks, which the recent incident around Tristan Buckmaster's use of Codex for solving the Navier-Stokes equation was involved (https://www.nytimes.com/2026/09/10/science/tristan-buckmaster-openai-math-navier-stokes.html). To put things on record, our President Abhivardhan had signed the Leiden Declaration on Artificial Intelligence and Mathematics as well (https://leidendeclaration.ai/signatories?q=Abhivardhan), representing the Indian Society of Artificial Intelligence and Law and OpenBharat.net.

5. We earnestly believe the Government of India should first strengthen the existing India AI Safety Institute with specific AI Safety and Security Research goals and child research groups. There were some vested interests pushing for private AI safety institutes, but to be honest - they are making the same mistake which many AI safety think tanks did - which is explictly pushing an agenda without a clear focus to solve problems beyond the "government tech" ecosystem. Private efforts on AI Safety in India should be guided by legitimate private interests and efforts. The sooner our policy think tanks realise this, the better. Or else, the government should at least strengthen IAIRO and IASI towards better directions.