Democrats Demand Congress Step In on AI Self-Regulation

Rep. Pete Aguilar says AI companies cannot be trusted to regulate themselves, so Congress has to step in with industry-wide rules for “public safety.” Fine. But we have heard this movie line before, and it usually ends with a thick rulebook written by staffers who cannot explain how the tool works, let alone how the rule will be enforced. “Congress must act” is easy to say. Building smart guardrails is the hard part.

The funny thing is Democrats talk like Washington has been a model regulator lately. It has not. The same crowd that cannot keep up with basic cyber threats, cannot pass a real privacy law, and cannot stop federal agencies from buying questionable tech now wants to referee the fastest-moving sector in the economy. That is how you get compliance theater: big firms lawyer up, smaller shops get crushed, and the public gets a false sense of security. A Washington rule that sounds tough can still be useless.

Yes, AI needs boundaries. If a company is training on stolen data, breaking the law, or putting out tools that enable real harms, hit them with real penalties. Use existing laws where they fit. Update the ones that do not. But “industry-wide rules” can become a backdoor wish list that has nothing to do with safety and everything to do with controlling speech, throttling open-source models, and picking winners. Don’t hand the keys to the same people who botched tech oversight for years.

If Democrats are serious, start with specifics. What harms, what standards, what agency, what timeline, what limits on bureaucratic sprawl? And make sure the rules do not turn into a gift to the biggest players who can afford an army of compliance officers. We can protect people without strangling innovation, but only if Congress stops grandstanding and starts doing the unglamorous work.