In a significant move aimed at bolstering cybersecurity, around 150 organizations in more than 15 countries, including Japan’s government agencies and financial institutions, have gained access to an advanced artificial intelligence system, Claude Mythos, developed by Anthropic. This initiative is designed for cybersecurity testing and research, allowing participants to leverage the AI’s capabilities in identifying system vulnerabilities and reinforcing defenses against cyber threats. Before participating, organizations must satisfy stringent security criteria.
Japanese Finance Minister Satsuki Katayama has confirmed the involvement of Japanese governmental bodies and banks, highlighting the initiative as a crucial step in facing future cybersecurity challenges. Claude Mythos is classified as a “frontier AI,” a term used for highly advanced systems that can detect weaknesses in complex digital systems. While these capabilities are invaluable for security research, experts caution that they could be exploited to compromise critical infrastructure if not properly managed.
Additionally, the European Union has also secured access to this technology. The European Union Agency for Cybersecurity (ENISA) is expected to be part of this program, with European officials expressing support for the initiative as part of their ongoing efforts to enhance cyber resilience across the region.
Amidst these developments, OpenAI is reportedly preparing to extend access to one of its own advanced AI models to major Japanese financial institutions. This move underscores the growing international interest in harnessing frontier AI technologies to strengthen cybersecurity defenses worldwide.