Overview
- Swiss regulator Marlene Amstad warned on Friday that AI is accelerating how quickly hackers find and exploit software flaws, and she urged banks to speed up patching and adopt new defenses.
- FINMA has helped create an IOSCO forum that represents watchdogs covering about 95% of global markets and convened roughly 100 policy and tech specialists for a hackathon to prototype AI tools for crypto and market supervision.
- Experience with advanced models such as Anthropic’s Mythos revealed high‑severity software vulnerabilities and prompted concern that the same tooling used for defense can also expose systems to faster attacks.
- The United States ordered Anthropic to suspend exports of its latest Mythos and Fable models this month, and Chinese firm 360 Security Technology said it has created a domestic alternative, reshaping which models are available for defensive use.
- Regulators are exploring embedding automated safeguards into digital‑asset systems and expanding supervisory technology but this will raise costs and compliance demands that could squeeze smaller crypto firms and change how banks deploy AI in customer services and operations.