On 5th of May 2026, FIIA hosted the public event entitled “Anthropic’s mythos model and the coming reality of AI-enabled cyber operations”.
The guest speaker was RAND’s Gregory Smith, who is a Visiting Research Fellow at FIIA for one month. The discussion was moderated by FIIA’s Leading Research Ville Sinkkonen, and the event was attended by policy makers, students, researchers and governmental experts.
The discussion focused on Anthropic’s recent announcement that its newest and most powerful model, Claude Mythos, was too dangerous to release to the public because of its ability to find and exploit critical vulnerabilities in existing critical software. Mythos indicates a future where AI becomes increasingly capable and influential across national security dimensions, from cybersecurity today to realms as diverse as biosecurity and core economic competitiveness tomorrow. The speaker discussed several implications, both short-term and long-term, for individuals and national governments alike. The active audience raised many questions ranging from strategic decision making in cyber operations to erosion of trust in democratic societies.
