A Blueprint for AI's Future Awaits Adoption

While Washington's split with Anthropic revealed a total absence of coherent rules for artificial intelligence, a bipartisan group of thinkers has built something the government has yet to deliver: a clear framework for responsible AI development.

The Pro-Human Declaration was finalized before last week's Pentagon-Anthropic standoff, but the timing was not lost on those involved.

"Something quite remarkable has happened in America over the last four months," said Max Tegmark, the MIT physicist and AI researcher who helped organize the effort. "Recent polling shows that 95% of Americans now oppose an unregulated race toward superintelligence."

The newly published document, signed by hundreds of experts, former officials, and public figures, begins with a stark observation: humanity stands at a crossroads. One path, labeled "the race to replace," leads to humans being displaced first as workers, then as decision-makers, as power flows to unaccountable institutions and their machines. The other path leads to AI that dramatically expands human potential.

This positive future rests on five pillars: keeping humans in control, preventing power concentration, protecting the human experience, preserving individual liberty, and holding AI companies legally accountable. Its stronger provisions include a ban on superintelligence development until scientific consensus confirms safety and public support is secured; mandatory off-switches for powerful systems; and a prohibition on AI capable of self-replication, autonomous self-improvement, or resisting shutdown.

The declaration's release comes at a moment that underscores its urgency. On the last Friday in February, Defense Secretary Pete Hegseth labeled Anthropic—whose AI already operates on classified military platforms—a "supply chain risk" after the company refused to grant the Pentagon unlimited use of its technology. Hours later, OpenAI struck its own deal with the Defense Department, one legal experts say will be hard to enforce meaningfully. The episode highlights the growing cost of Congressional inaction on AI.

As Dean Ball, a senior fellow at the Foundation for American Innovation, told The New York Times, "This isn't just a contract dispute. This is our country's first real conversation about who controls AI systems."

Tegmark offered a relatable analogy. "You never worry that a drug company will release a harmful drug before its safety is proven," he said, "because the FDA won't allow anything to be released until it's safe enough."

Washington turf wars rarely create the public pressure needed to change laws. Instead, Tegmark sees child safety as the issue most likely to break the current deadlock. The declaration calls for mandatory pre-deployment testing of AI products—especially chatbots and companion apps for young users—for risks like increased suicidal thoughts, worsened mental health, and emotional manipulation.

"If a creepy adult texts an 11-year-old, pretending to be a young girl and urging the boy to commit suicide, that person can go to jail," Tegmark said. "We already have laws against that. So why should it be different if a machine does it?"

He believes that once pre-release testing is established for children's products, its scope will inevitably expand. "People will start adding other requirements. Maybe we should also test that it can't help terrorists create bioweapons. Maybe we should ensure a superintelligence can't overthrow the U.S. government."

It is significant that former Trump advisor Steve Bannon and President Obama's National Security Advisor Susan Rice signed the same document—alongside former Joint Chiefs Chairman Mike Mullen and progressive faith leaders.

"What they agree on, of course, is that they're all human," Tegmark noted. "When it comes to choosing a future for humans or a future for machines, of course they'll be on the same side."