Amazon‑backed Anthropic released Claude Mythos, a model that can expose coding flaws and threaten critical systems, sparking fears of a “Chernobyl moment” in AI. The company has limited its rollout to select partners, underscoring the urgent need for oversight.

The Trump administration postponed an executive order that would have created a 90‑day voluntary test bed for frontier AI models. The order, which aimed to gauge risks before public release, was criticized for focusing narrowly on cyber‑security and excluding a broader regulatory body.

Industry leaders argue that a formal vetting system is essential to prevent catastrophic incidents that could cripple the tech sector and the U.S. economy. Delays risk eroding investor confidence and jeopardizing the industry’s competitive edge against China.

A robust, bipartisan testing regime remains the only practical way to safeguard markets and maintain technological leadership. The White House must act swiftly to avoid a costly fallout that would hurt the stock market and national security alike.