en.Wedoany.com Reported - Anthropic is opening access to its cybersecurity AI system, Mythos, to the European Union Agency for Cybersecurity (ENISA) and incorporating it into the Project Glasswing initiative. Previously, access to the system was primarily available to US partners, including government agencies, enterprises, and multiple banks. The UK's Artificial Intelligence Security Institute has also been authorized to test Mythos, aiming to leverage it to protect critical software security before similar capabilities become widely available. ENISA is the first EU institution to gain access to Mythos.
An ENISA spokesperson confirmed to the UK's Financial Times that negotiations regarding Mythos access have taken place, but emphasized that the terms are still under discussion. According to sources familiar with the matter, the negotiations also cover the scope of access for the US company Anthropic to EU systems when European institutions use Mythos.
The European Commission has also confirmed negotiations with Anthropic. A spokesperson stated to Bloomberg that the Commission has held several productive meetings with the company and welcomed recent progress in opening access, though no specific timeline was disclosed. Related reports indicate that European Commission representatives traveled to San Francisco last week to negotiate with Anthropic regarding joining the Project Glasswing initiative.
Since the launch of Mythos in April this year, the EU has been seeking access to it but had long been unsuccessful. During the limited opening, Anthropic repeatedly emphasized that the model possesses extraordinary capabilities for discovering security vulnerabilities, along with associated potential for misuse.
Anthropic stated last week that it would pursue broader access, while also mentioning models at the Mythos level. According to the company, the Claude Mythos Preview version currently being tested under the Project Glasswing framework will not be made available to the public.
Meanwhile, European banks are seeking their own pathways to address cybersecurity AI threats. For example, BNP Paribas has expanded its partnership with Paris-based AI startup Mistral AI and is preparing for a new generation of powerful cybersecurity AI models. Even if the EU gains access to Mythos, the risks posed by AI-driven cyberattacks will not be eliminated as a result.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
