Thanks for quoting me CUTS Institute for Regulation and Competition (CIRC)!
Last week, I had a fantastic session alongside Taiwanese, Chinese and American stakeholders discussing some hard questions around AI, Geoeconomics and Geopolitics.
Complete session
My position was simple:
1️⃣ LLM valuations are artificially inflated due to data pipeline issues and modest contextual assistance being oversold as revolutionary capability, setting the stage for market correction.
2️⃣ India hasn’t fallen behind in AI because current LLMs are architecturally designed with performance caps that restrict them to mundane applications rather than breakthrough innovation.
3️⃣ Despite 400+ AI policy documents since Bletchley 2023, the decentralized global AI community has failed to create actionable governance frameworks, leaving comprehensive AI oversight virtually non-existent. Do check my friend Var Shankar‘s contributions on end-to-end AI governance to World Economic Forum‘s latest AI governance playbook as well.
The LLM ecosystem is really subject to huge saturation, and that will create market and talent credibility challenges.
My point was simple: the AI community is globally decentralised and regulatory arbitrage around AI might just cool off due to LLM tech failures.
So has India lost the AI race? I don’t think so yet. While “Sovereign AI” initiatives should continue, I don’t think we have lost. But dynamics will be rapid once the LLM “banner” falls off.
However, investors have made a lot around data outflows, for sure thanks to data labelling and other associated practices.
Also, a special shoutout to ExtensityAI - and Sankalp Srivastava - because our lovely report on AI & Market Power in a Multipolar World was referred to by AI Austria :)
