Indian Fintech Leaders Seek Early Access to Anthropic's High-Risk 'Mythos' AI

Leading financial technology companies in India are urging Anthropic PBC to grant them early access to Mythos, the developer's latest large-language model. The push is driven by a need to proactively secure financial infrastructure against potential AI-driven threats. The request comes from prominent industry players, including One97 Communications Ltd., Razorpay Software Ltd., and Pine Labs Ltd. These firms aim to use the model to conduct stress tests and detect vulnerabilities within their own systems before the technology becomes more widely available. Anthropic has maintained a restricted rollout for Mythos, citing concerns that the model is too dangerous for a general release. The AI community has expressed fears that the model's capabilities could usher in a new era of highly sophisticated cyberattacks, particularly targeting sensitive financial data. While no breaches have been reported, the urgency from India's fintech sector highlights the growing arms race between AI developers and cybersecurity specialists. The ability of financial institutions to 'red-team' their systems using the same tools potential attackers might use is becoming a critical component of modern risk management.