Speaker Johnson: AI Leaders Must Discuss Safety, No Moratorium

sit the AI bosses down with lawmakers and the White House, and write rules that actually work. Not a commission that meets for a year and produces a glossy PDF. Not another round of “listening sessions” where everyone nods and nothing happens. A serious industry needs serious guardrails.

The other part matters just as much: no moratorium. You can’t “pause” your way to safety when China isn’t pausing, and when the next breakthrough will come from whoever keeps building. America has to lead because we’re the ones who can demand standards, verify compliance, and enforce consequences. We don’t win by stepping aside. If Washington is worried about deepfakes, job disruption, cyber threats, and AI tools in the hands of criminals, then the answer is clear rules and real enforcement, not freezing the field and hoping the world follows.

But Congress can’t treat “safety” like a magic word that covers everything. Safety has to mean specific things: security testing for models, protections for critical infrastructure, transparency around training data where it matters, and liability when companies ship reckless products. And yes, the federal government has to get out of the habit of writing laws that only Big Tech can afford to comply with. Don’t turn “AI safety” into a monopoly plan.

Johnson’s push is a good start because it recognizes the basic reality: this is a strategic competition and a domestic governance problem at the same time. The grown-up move is to keep building, keep leading, and set rules that punish bad actors without smothering the inventors who can keep America ahead. That’s not hype. That’s the job.