The world must “get ready” for powerful AIs that can exploit vulnerabilities in websites within the next few months, Anthropic’s co-founder and head of public benefit warned.

Claude Mythos, Anthropic’s new AI model, sparked alarm at the company when it discovered previously unknown security flaws in every major web browser and operating system. Anthropic has only released it to select organizations so that they can help Anthropic plug any further holes.

Speaking at Semafor World Economy, Jack Clark said that Mythos is “not a special model” and that equally capable ones will be coming from other companies soon, and within a year and a half, “there’ll be open-source models from China that have these capabilities.”