AI Planning and Design

Sometimes the great verities are pretty straightforward. For example, “If something cannot go on forever, it will stop,” an observation made famous by economist Herb Stein. Let’s stick with that one, given its sensible application to the Trump administration’s current quasi-licensing regime for frontier AI models. The effort has proven to be a vibey, ad hoc process that The Economist correctly describes as “opaque, unpredictable—and unsustainable.” Good news that a more structured process apparently is coming soon.

The Financial Times reports that, as early as this week, the White House will roll out “voluntary standards for the release of new models” that would “set benchmarks for models with cutting-edge cyber capabilities and establish release timelines in an effort to streamline future launches.” What’s more, the guidelines would clarify “who is able to access advanced models, both domestically and abroad, in a move that could set the stage for a global framework including US allies.” Definitely seems less haphazard, I’ll give it that.

Late last week, Anthropic took its new Claude Fable 5 and Mythos 5 AI models offline following a United States government export-control directive barring “any foreign national” from using the services. The company has been in talks with the White House since Friday but has yet to secure an agreement that would allow it to reinstate the offerings.

Since Mythos debuted in April, Anthropic has claimed—and warned—that the model has advanced capabilities for not only finding software vulnerabilities to help defenders patch them, but also figuring out ways to exploit them that could be used by bad actors. Anthropic itself noted this double edged sword in its launch of Mythos 5 and Claude Fable 5. “A great deal of advanced usage of AI models is dual use: the same queries that are beneficial in the hands of cybersecurity professionals and biology researchers could be dangerous if available to malicious actors,” the company wrote in a blog post last week.

Anthropic is backtracking on a policy that would have covertly limited competitors from using its new AI model, Claude Fable 5, to develop other AI models. The company changed course after the move received significant backlash from the AI research community.

“We’re changing Fable 5’s safeguards for frontier LLM development to make them visible,” Anthropic said in a statement to WIRED. “We made the wrong trade-off and we apologize for not getting the balance right.”

Anthropic released Claude Fable 5, a version of its latest AI model with additional safety guardrails designed to prevent misuse, earlier this week. Some of the safeguards Anthropic decided on were unsurprising: The company said it would reroute users who asked questions about cybersecurity, biology, or chemistry to a less capable AI model to reduce the chances of someone using the advanced AI to carry out a cyberattack or build a bioweapon.

The buzziest bit of the new artificial-intelligence bill from Reps. Jay Obernolte, a California Republican, and Lori Trahan, a Massachusetts Democrat, is probably the section that would let Washington preempt state rules on AI development for three years. For me, the more interesting part is its bet on auditing as a middle path between Silicon Valley self-regulation and an FDA-style premarket approval regime for frontier models.

Earlier this year, several dozen AI policy folks signed onto a proposal, “Frontier AI Auditing: Toward Rigorous Third-Party Assessment of Safety and Security Practices at Leading AI Companies” that “outlines a vision for frontier AI auditing, which we define as rigorous third-party verification of frontier AI developers’ safety and security claims, and evaluation of their systems and practices against relevant standards, based on deep, secure access to non-public information.”

What does that mean? For starters, an audit is not a permission slip. Rather than making a company clear a government gate before shipping, as the FDA does with drugs, an independent reviewer with access to the confidential innards of an AI company would check the latest model against a set standard. Think of it as how an accountant signs off on a public company’s books. Private firms do the examining while a public body stays in the background. Maybe it sets the rules, accredits the examiners, and holds the enforcement hammer.