Anthropic eyes major Claude Mythos expansion as India flags bank security risks

Anthropic wants more companies to use Mythos but the White House is saying no. Mythos is an all-new version of Claude made specifically for warding off cyberattacks. It is apparently so powerful that it is only available to select users part of a super-elite Project Glasswing. Currently this number is 40 plus and includes companies like Apple, Google and Microsoft. But reports suggest that Anthropic wants to expand and make Mythos available to 70 more clients.

According to The Wall Street Journal, the White House is closely involved in how Mythos is being rolled out. This is because Claude Mythos, given how powerful it is, is under the radar for potential national security risks. Globally, too, countries and governments are watching Mythos closely, practicing caution particularly in banking and finance. The reasoning is that Mythos is designed to find loopholes and vulnerabilities in software with the intension that if it can find something off, it can be fixed before somebody exploits it. The fear and concern is that it could be a dangerous weapon if it falls in the wrong hands.

For Anthropic, limited use would mean that Mythos would have limited opportunities to be more efficient and productive. This is because AI models only get better with training and deployment. The US government is pushing back over security concerns.

Tensions remain between Anthropic and the US government

This decision also reflects broader tensions between Anthropic and the US government. The White House’s refusal shows that the relationship is still complicated, especially as the administration is currently involved in two separate court cases with Anthropic after it declined to allow its AI tools to be used for fully autonomous weapons systems.

Unauthorised access under investigation

Even as access remains restricted, there are fresh concerns around control. Anthropic does not plan a public rollout yet, but a small group of unauthorised users recently gained access to Mythos. The company is aware of the breach and is now investigating the matter.

India steps in

At the same time, global interest in Mythos continues to grow. Governments and organisations worldwide are preparing for a wider rollout, and India is also closely watching developments. Finance Minister Nirmala Sitharaman recently held a high-level meeting with top bank officials, where discussions focused on the potential cybersecurity risks linked to the model.

Building on this, the Indian government has stepped up talks with the United States and Anthropic. It is exploring ways to ensure Indian companies get fair access to the model while also protecting critical systems such as banking networks, telecom infrastructure, and power grids. Officials are said to be working on both “mechanisms” and “logistics” as part of ongoing bilateral discussions with the US administration.