Here's the unpopular take: restraint, not speed, may be the smarter strategy here.
As Washington scrambles to regulate artificial intelligence, the momentum is undeniable. Kill switches, licensing requirements, systemic risk frameworks, data center permitting overhauls. The pressure to act decisively has never felt more urgent. Tech companies are pushing for clarity. Lawmakers are pushing for control. Everyone agrees we need rules.
But faster regulation isn't always better regulation. And right now, the velocity of Washington's AI oversight push risks producing rules that are simultaneously too rigid and too fragile.
Consider the structural problem we're facing. Five years ago, nobody in Congress could articulate what a large language model was. Today, they're drafting bills that would govern the frontier of AI development with the kind of specificity usually reserved for nuclear power plants. The knowledge gap hasn't closed. It's just been rushed over.
When you move this quickly on a technology that changes this fast, you tend to regulate yesterday's version of the problem. You lock in today's assumptions as law. You accidentally protect incumbent firms by raising barriers to entry that only they can afford to clear. You write language that seems precise but becomes obsolete before the ink dries.
The recent headlines tell this story implicitly. A Treasury analysis flagged AI as a systemic risk, then Treasury disowned it. That reversal wasn't necessarily wrong. It was honest about how little consensus exists on what systemic risk even means in this context. Yet that uncertainty hasn't slowed the legislative pace. It's just created a vacuum where companies and agencies are all trying to regulate simultaneously, often at cross purposes.
Washington wants to price frontier AI development. Companies want to expand data centers. Congress wants kill switches. Regulators want to identify systemic risks they can't quite define. All of this is happening at once, and none of it is clearly coordinated.
The pressure to move fast creates perverse incentives. Industry players propose regulations that benefit them while appearing reasonable. Lawmakers chase the appearance of action without fully understanding the consequences. Everyone rushes because nobody wants to be blamed when something goes wrong. But speed in this environment produces rules that satisfy nobody and may actually increase the risks they're designed to prevent.
What would restraint look like instead?
It would mean establishing clearer baselines of understanding before writing prescriptive rules. It would mean distinguishing between what actually needs federal legislation immediately versus what could be managed through agency guidance, industry standards, or sectoral approaches. It would mean acknowledging that some problems aren't solvable through regulation alone.
It would mean resisting the temptation to regulate every potential future scenario. And it would mean being honest that we're making decisions with genuine uncertainty, which argues for flexibility and revision built into whatever frameworks we create.
None of this means doing nothing. The current moment calls for governance, oversight, and accountability. But it calls for calibrated governance. Measured oversight. Accountability mechanisms that can actually work rather than those that look good in a press release.
The risk of moving too slowly on AI regulation is real. The risk of moving too fast and locking in the wrong rules is equally real. We don't have a good track record as a country of undoing bad regulations once they're written.
So here's what genuine urgency would look like: taking the time to get it right. Building in review periods. Creating pathways to adapt rules as the technology and our understanding of it evolve. Coordinating across agencies and levels of government rather than racing to regulate first.
It's not sexy. It won't produce headlines about decisive action. But it might actually produce regulations that serve the public interest rather than just appearing to do so.