Why the AI Industry Needs a Clear Roadmap — and Who Should Be Leading It

As Washington's fractured relationship with Anthropic laid bare the absence of any coherent regulatory framework governing artificial intelligence, a bipartisan coalition of researchers, former officials, and public intellectuals has done what the federal government has so far refused to: articulate a concrete vision for what responsible AI development should actually look like.

The Pro-Human Declaration was finalized before last week's Pentagon-Anthropic standoff, but the collision of the two events was hardly lost on those who helped bring it into existence.

"There's something quite remarkable that has happened in America just in the last four months," said Max Tegmark, the MIT physicist and AI researcher who helped organize the effort, in conversation with this editor. "Polling suddenly [is showing] that 95% of all Americans oppose an unregulated race to superintelligence."

The newly published document, signed by hundreds of researchers, former senior officials, and prominent public figures, opens with a stark observation: humanity is standing at a genuine inflection point. One trajectory — which the declaration terms "the race to replace" — leads to a world where humans are progressively displaced, first as workers, then as decision-makers, as institutional and technological power consolidates beyond any meaningful democratic accountability. The alternative envisions AI as a force multiplier for human capability, augmenting rather than supplanting the people it serves.

That second path, the declaration argues, rests on five foundational pillars: preserving meaningful human oversight, preventing dangerous concentrations of power, safeguarding the integrity of human experience, protecting individual liberty, and establishing enforceable legal accountability for AI developers. Among its most substantive provisions is an outright moratorium on superintelligence development until scientific consensus confirms it can be pursued safely and with genuine democratic legitimacy. The document also calls for mandatory kill-switch mechanisms on high-capability systems and an explicit prohibition on architectures capable of autonomous self-replication, recursive self-improvement, or active resistance to shutdown — technical constraints that align closely with what AI safety researchers have long identified as core risk vectors.

The declaration's release arrives at a moment that sharpens its urgency considerably. On the last Friday in February, Defense Secretary Pete Hegseth designated Anthropic — whose AI systems already operate on classified military infrastructure — a "supply chain risk" after the company declined to grant the Pentagon unrestricted access to its technology. It is a designation typically reserved for firms with documented ties to adversarial foreign governments. Within hours, OpenAI reached its own agreement with the Defense Department, one that legal analysts say will prove difficult to enforce in any operationally meaningful way. Taken together, the episode exposed just how costly years of congressional inaction on AI governance have become.

As Dean Ball, a senior fellow at the Foundation for American Innovation, told The New York Times in the aftermath, "This is not just some dispute over a contract. This is the first conversation we have had as a country about control over AI systems."

Tegmark reached for a regulatory analogy that cuts through the abstraction. "You never have to worry that some drug company is going to release some other drug that causes massive harm before people have figured out how to make it safe," he said, "because the FDA won't allow them to release anything until it's safe enough." The parallel is deliberate: in virtually every other high-stakes industry, pre-market safety validation is not a courtesy — it is a legal prerequisite. AI, so far, operates under no equivalent obligation.

Washington turf battles rarely generate the kind of sustained public pressure that translates into legislative action. Tegmark, however, sees child safety as the political pressure point most likely to break the current regulatory deadlock. The declaration reflects this calculus directly, calling for mandatory pre-deployment testing of AI products — with particular scrutiny applied to chatbots and companion applications targeting younger users — covering documented harms including elevated suicidal ideation, exacerbation of underlying mental health conditions, and algorithmic emotional manipulation.

"If some creepy old man is texting an 11-year-old pretending to be a young girl and trying to persuade this boy to commit suicide, the guy can go to jail for that," Tegmark said. "We already have laws. It's illegal. So why is it different if a machine does it?"

His theory of regulatory momentum is incremental by design. Once the principle of mandatory pre-release evaluation is established for consumer-facing products aimed at children, he argues, the scope of required testing will expand almost organically. "People will come along and be like — let's add a few other requirements. Maybe we should also test that this can't help terrorists make bioweapons. Maybe we should test to make sure that superintelligence doesn't have the ability to overthrow the U.S. government." It is a legislative strategy that mirrors how consumer protection law has historically evolved: one category of harm at a time, until the framework becomes comprehensive.

Perhaps the most striking feature of the declaration is the ideological breadth of its signatories. That former Trump advisor Steve Bannon and Susan Rice, President Obama's National Security Advisor, have affixed their names to the same document — alongside former Joint Chiefs Chairman Mike Mullen and progressive faith leaders — is not a trivial detail. It signals that concerns about AI's trajectory have begun to transcend the conventional fault lines of American political life.

"What they agree on, of course, is that they're all human," says Tegmark. "If it's going to come down to whether we want a future for humans or a future for machines, of course they're going to be on the same side."