India Warns of Cyber Risks as Anthropic's Mythos AI Breaches Defenses in 83 of 100 Test Cases

According to The Economic Times, India’s cybersecurity agency recently issued a high-severity advisory after Anthropic’s Mythos AI demonstrated the ability to find tens of thousands of vulnerabilities and convert software bugs into exploitable attacks in minutes.

Palo Alto Networks’ Unit 42 and Check Point found Mythos discovered tens of thousands of vulnerabilities compared to approximately 500 in Anthropic’s earlier Opus 4.6 model, breaching defenses on the first attempt in 83 of 100 test cases. The advisory highlighted that while Indian firms typically require 60 to 90 days to patch systems, Mythos can turn bugs into attacks in minutes. India’s Ministry of Electronics and Information Technology is reportedly discussing early access to the model amid concerns over risks to banking, telecom, power, and rail systems.