NEW DELHI — Following the emergence of Anthropic's 'Mythos Preview' AI model, which autonomously developed 181 working exploits and achieved register control on 29 more in April 2026, Indian regulators are re-evaluating their 'light touch' approach to artificial intelligence. Previously, in 2025, Anthropic's Opus 4.6 model had a 'near-0% success rate' at autonomous exploit development, leading officials to believe minimal oversight was sufficient for rapid economic incentivization.

"We did not explicitly train Mythos Preview to have these capabilities. Rather, they emerged as a downstream consequence of general improvements in code, reasoning, and autonomy," explained a spokesperson for Anthropic, referring to their April 7 note. This unexpected self-improvement has forced the Indian financial sector to be told to exercise a 'high-degree' of vigilance and develop coordination mechanisms, a process known for its lightning speed.

The rapid, exponential increase in AI's offensive cyber capabilities, far outstripping previous generations, has caught regulators off-guard, despite prior warnings about AI risks. An unnamed official noted that while India aims for a 'seat at the high table' in global AI governance, the immediate priority is understanding how a technology whose creators admit they don't fully control managed to bypass the nation’s celebrated 'light touch' regulatory framework.

"We are monitoring the situation closely," stated a government representative, echoing sentiments from every previous technological disruption that initially promised economic growth before raising unforeseen national security concerns.