Artificial intelligence has officially entered the “Congress is trying to do something about it” phase, which means we’ve now moved from “terrifyingly fast technological disruption” to “terrifyingly fast technological disruption, but with committee hearings.”

According to reports surrounding the Obernolte-Trahan AI negotiations, Reps. Jay Obernolte, a California Republican, and Lori Trahan, a Massachusetts Democrat, have been working on a bipartisan framework that could temporarily preempt some state AI laws, especially those dealing with AI model development. The basic idea appears to be this: instead of letting California, New York, Colorado, and every other state build their own AI rulebook, Congress would create a national framework, possibly with a two-year pause on certain state-level regulations while federal standards are developed.

This isn’t just about “regulation” in the abstract. It’s about who gets to set the rules for one of the most powerful technologies ever released into public life: Congress, states, federal agencies, courts, industry, or some cheerful combination of all of the above that will definitely not become a bureaucratic casserole.

Why a National Standard Could Help

The strongest case for federal AI legislation is simple: America needs one serious national framework, not fifty different experiments duct-taped together.

AI doesn’t respect state borders. A model trained in California, deployed by a company in Texas, used by a school in Delaware, and hosted on servers somewhere near a power plant in Virginia isn’t exactly easy to regulate one ZIP code at a time. A patchwork approach may sound empowering in theory, but in practice it could create confusion, compliance nightmares, and an advantage for the biggest companies that can afford armies of lawyers.

That last point matters. We shouldn’t assume that more state regulation automatically means more accountability. Sometimes it means small businesses get crushed while Big Tech shrugs and says, “Good thing we already have a regulatory affairs department the size of a small nation.”

A national framework could also help the United States compete globally. The American Leadership in AI Act attempts to consolidate more than 20 bipartisan proposals into one broader package addressing AI standards, evaluations, research infrastructure, federal procurement, worker issues, deepfakes, and AI education. That shows there’s at least some bipartisan recognition that AI policy can’t remain a food fight forever.

There’s also a national security argument. China isn’t waiting for California, New York, and Congress to finish a group project. If the U.S. buries AI development under a tangled mess of inconsistent rules, we may slow down our own innovators while adversaries press ahead. That doesn’t mean “let the robots do whatever they want,” but it does mean regulation should be careful, targeted, and realistic.

This is where prudence matters. Government has a legitimate role in punishing wrongdoing, protecting the innocent, and maintaining public order. But regulation shouldn’t become a reflexive lurch toward control every time something new appears. Fire is dangerous, but civilization didn’t advance by banning fireplaces.

Why States Shouldn’t Be Sidelined Too Quickly

The strongest argument against the bill is that federal preemption could become a polite way of saying: “States, please sit down while Washington writes a weak law the tech industry can live with.”

That concern isn’t imaginary. AI safety advocates are worried that preempting state laws before Congress has enacted meaningful protections would remove the strongest existing pressure on AI companies. The Washington Post reported that safety advocates fear the Obernolte-Trahan discussions could hand too much leverage to the pro-industry camp, especially if Congress blocks future state laws while offering only a limited federal oversight regime in return.

That’s a serious concern. States often function as laboratories of democracy. When Washington is slow, distracted, or busy arguing about whether the vending machines in the Capitol are ideologically biased, states can act. California and New York may overreach at times, because of course they can. But state-level action can also expose real harms, force transparency, and create pressure for national standards.

The danger is especially obvious with children, privacy, deepfakes, fraud, and civil rights. AI systems can manipulate vulnerable users, generate sexualized or defamatory images, impersonate real people, influence hiring or lending decisions, and reshape education faster than many parents and teachers can understand what’s happening. This isn’t just “innovation.” It’s power.

And conservatives, of all people, should be skeptical when powerful corporations beg Washington to save them from local accountability. Big Tech hasn’t exactly earned the public’s blind trust. These are the same people who gave us algorithmic addiction machines, political censorship controversies, and creepy data harvesting.

So, when the tech industry says, “Trust us, we just need regulatory certainty,” the proper response isn’t hostility, but it’s also not childlike wonder. It should be: “Wonderful. Show us the text, the enforcement mechanism, the liability rules, the child protections, the privacy standards, and the campaign donations. In no particular order.”

Why the Tradeoffs Deserve a Serious Look

Both sides have a point, and both sides have incentives they would rather not advertise.

Supporters of federal preemption are right that AI needs national coordination. A fragmented regulatory system could become chaotic and expensive, especially for smaller firms, schools, nonprofits, and local businesses trying to use AI responsibly. A serious federal framework could provide clarity, strengthen U.S. competitiveness, and prevent states from creating rules that are more symbolic than workable.

Critics are right that preemption without strong federal safeguards would be reckless. If Congress pauses state laws and then replaces them with voluntary guidelines, vague standards, or underfunded enforcement, that’s not regulation.

The central question isn’t whether federal law should exist. It should. The question is whether federal law becomes a floor or a ceiling. A floor establishes minimum protections while allowing states to go further in targeted areas. A ceiling blocks states from doing more, even if new harms emerge and Congress goes back into its natural resting state: fundraising and blaming the other party.

That distinction is everything.

Why Human Dignity Must Stay at the Center

AI should be judged by a few basic principles.

First, human dignity must come before technological dazzlement. Man is made in the image of God; machines are not. AI may be useful, impressive, and economically transformative, but it must remain a tool under human moral responsibility. When AI affects children, families, employment, speech, privacy, education, criminal justice, or medical care, government shouldn’t pretend the only moral question is whether quarterly earnings look nice.

Second, innovation is good, but not ultimate. Christians shouldn’t fear human creativity. Building, discovering, designing, and improving are part of mankind’s stewardship calling. But innovation severed from wisdom becomes Babel with better branding. “Can we?” must never replace “Should we?”

Third, limited government doesn’t mean absent government. A conservative approach should reject both extremes: a suffocating regulatory state that strangles progress, and a libertarian fantasy in which giant AI companies police themselves out of the goodness of their shareholder reports. Government should be limited, constitutional, accountable, and competent. Yes, that last one is ambitious.

Fourth, parents and local communities matter. Any AI framework that weakens protections for children, undermines parental authority, or leaves families defenseless against addictive and manipulative systems should be opposed. Children aren’t beta testers for trillion-dollar experiments.

Support the Goal, Don’t Trust the Shortcut

My verdict: Congress should pursue federal AI legislation, but we should be very cautious about any bill that preempts state laws before strong federal protections are actually in place.

A national AI framework is necessary. The country can’t regulate this technology effectively through fifty separate rulebooks forever. But federal preemption should be earned, not handed out like a party favor at a tech lobbyist reception.

The right approach would be a federal floor: clear national standards for transparency, child safety, privacy, deepfakes, fraud, high-risk uses, and accountability, with real enforcement and liability where harm occurs. Only after that should Congress consider limiting conflicting state laws. And even then, states should retain room to address emerging harms, especially when Congress inevitably discovers that “moving fast” isn’t one of its spiritual gifts.

So, yes to thoughtful federal AI legislation; yes to American innovation; yes to protecting children, families, workers, and civil society; no to giving Big Tech a preemption shield in exchange for vague promises and glossy white papers.

AI may be new, but the political temptation is ancient: powerful interests want freedom for themselves and rules for everyone else. We should be wise enough not to confuse that with liberty.