Anthropic offers the EU access to its Mythos cyber security AI model

In short: Anthropic has offered the European Union access to Mythos, its AI tool for finding and fixing cyber security weaknesses, and EU officials are negotiating the terms.

What happened

Anthropic, a US artificial intelligence company, has offered the EU access to its Mythos model. This is the first time the company has made the tool available outside the US and the UK.

An EU cyber security agency spokesperson confirmed that Enisa, the EU agency that helps countries coordinate on cyber threats, is in talks to use Mythos. The spokesperson said it has been offered, but the conditions are still being agreed.

European Commission officials visited San Francisco last week to discuss joining Project Glasswing. This is a group of mostly US companies that have been using Mythos since early April to find and patch security vulnerabilities, which are like weak spots in software that attackers can exploit.

Some of the key details are still unsettled. People familiar with the discussions said the EU and Anthropic still need to define how access would work, including how much access Anthropic might get to EU systems while the tool is being used.

Why it matters

Cyber attacks can disrupt hospitals, banks, and government services, and they can expose personal data. Tools like Mythos are meant to help defenders spot problems faster, like using a very skilled inspector to look for cracks in a building before it collapses. But the talks also show a tradeoff, since using an outside company’s tool can raise questions about control, oversight, and what information the company might be able to see.

Source: Financial Times