Anthropic Mythos is scaring banks, Big Tech

Heads of the Federal Reserve and Treasury convened an emergency discussion with the leaders of major American banks on Friday to discuss the destructive potential of Anthropic’s Mythos AI model.

The meeting attendees included CEOs of Bank of America, Citigroup, Goldman Sachs, Morgan Stanley, and Wells Fargo.

The Mythos is capable of detecting cybersecurity flaws exponentially faster and more efficiently than humans. Anthropic admits that the new model represents a significant leap in capability, having already identified “thousands of high-severity vulnerabilities, including some in every major operating system and web browser.”

Because of these risks, Anthropic has limited distribution to approximately 40 organizations.

CEO Dario Amodei stated that the company has also launched Project Glasswing to mitigate these risks, partnering with tech giants, such as Apple, Google, Microsoft, and Nvidia, to harden digital defenses.

Amodei maintains that while the dangers of getting Mythos wrong are “obvious,” the model offers a unique opportunity to build a fundamentally more secure internet.

Meanwhile, the International Monetary Fund (IMF) Managing Director, Kristalina Georgieva, expressed deep concern over the exponential growth of AI and the cyber risks it poses, specifically highlighting the Mythos model.

She warned that the international monetary system currently lacks the necessary defenses to thwart massive AI-driven attacks.