US lawmakers split on AI regulation amid safety fears

On the floor of a fulfillment center outside Columbus, Ohio, a woman named Loretta picks orders for eleven hours a shift, and the pace she walks is not set by her supervisor. It is set by a piece of software that scores every second she is not moving toward a shelf. That system has already spun out of anyone’s control but the company that owns it. The debate happening in Washington this week will decide whether anyone in government even has the authority to ask that company a question about it.

Congress came back from recess on Monday to warnings that artificial intelligence could become dangerously uncontrolled, and by every account they went home for lunch no closer to agreement on what, if anything, to do. The story is being told as a story about the future - about systems that might someday outthink the people who built them, might someday make decisions no human can reverse. That is a real fear and I will not mock it. But notice what the fear is not about. It is not about Loretta’s picking rate. It is not about the algorithm that decides which warehouse worker gets flagged for “low productivity” and quietly managed out before she qualifies for a pension. The thing that is actually, presently, measurably out of control in this country is not some future superintelligence. It is the authority companies have already seized, today, in 2026, to let a machine set the terms of a person’s working life with no appeal and no explanation.

Here is the arrangement, and it is not an accident. When the conversation in Washington is pitched at the level of civilizational risk - AI that might deceive us, might replace human judgment entirely, might threaten national security - every actor with a stake in deployment gets to sound responsible simply by nodding along to the danger. The company that owns Loretta’s scheduling algorithm can testify before a Senate committee about existential risk with a straight face, because that conversation costs it nothing. It is the conversation about the pace-setting software running in its own warehouses right now that would cost it something, and that conversation is nowhere on the agenda I have seen. The stakes described this week - national security, the economy, the public - are real, but they are pitched at an altitude where nobody has to look down and see the shop floor at all.

This is the oldest trick in the ledger book. Give the alarmed public a monster in the distance, vast and vague, and they will spend their attention bracing for it while the smaller, nearer thing walks in the front door unchallenged. I have seen mine owners do this with safety inspections - talk gravely about some hypothetical catastrophic collapse while the roof over the current shift was already rotten. The distant danger becomes the permission structure for ignoring the immediate one.

The honest question the members debating this in Washington ought to be asking is not only whether AI might someday spin out of control. It is who controls it right now, over whom, and under what appeal. Ask the fulfillment center company what happens when Loretta disputes her productivity score. Ask the call center what recourse a worker has when an AI system scores her tone of voice and marks her for discipline. If Congress cannot answer who holds power over a machine’s decisions in a warehouse in Ohio today, I do not believe it is equipped to answer who will hold power over a smarter machine tomorrow. The muscle for oversight is built on the small cases or it is not built at all.

So when you read that Washington lacks agreement on how to intervene, do not just picture senators arguing about the far horizon. Picture Loretta, watching a number on a handheld scanner tell her whether she still has a job at the end of the month, with nobody in that chamber asking her name.