Hermer demands urgent law to control artificial intelligence

The comparison is tidy. Richard Hermer says artificial intelligence faces less regulation than a refrigerator, and concludes that we need urgent public debate and robust international law. One notices the refrigerator. It was chosen as the punchline, but it is the most useful object in the argument. Once you treat it as load-bearing, the remedy Hermer proposes looks like the opposite of what the refrigerator teaches.

Consider how a refrigerator actually came to be regulated. There was no treaty conference, no summit communique and no public debate about the future of cold. Engineers wrote down what the machine must not do: shock the user, leak its refrigerant, catch fire behind the kitchen units, or lie about its energy use on the label. Each of those failures can be observed in a test lab by a person with a clipboard, a multimeter and a free afternoon. The rules are dull and testable, and that dullness is why they work. They also became international without diplomats. The household appliance safety standard, IEC 60335, travels across borders because a test written in Geneva produces the same result in Rotterdam and Shenzhen. Nobody had to agree on the meaning of refrigeration. They had to agree on what counts as a fail.

The strongest objection deserves its full weight. A refrigerator has a fixed purpose and fixed physics. A frontier model is general-purpose, changes with every release, and is built by a handful of firms whose compute sits in several jurisdictions. A national product standard is a fence around one field when the animal grazes on four continents, so international law looks less like a preference and more like a necessity. The objection also points out that Europe already has the AI Act, so the fridge line is not quite true everywhere, and wherever it is true the cause is the reach of borders.

The objection is right about the borders and wrong about the order of operations. International law never creates the test. It ratifies one that already works. Arms control treaties could count warheads because someone could already count warheads. A treaty that bans losing control of the future, which is the stake Hermer names, has nothing to count. It would turn on a phrase no inspector can measure, and its enforcement would be done by press release. The Dutch call the posture that results de kat uit de boom kijken, watching the cat out of the tree: everyone stays still and waits to see which way the thing jumps before committing. Public debate tends to produce exactly that.

Here is the plain question for Hermer and for anyone nodding along. What is the test? Not the principle or the value, but the procedure a technician runs on a Tuesday that ends in pass or fail. If the answer is that we will work it out through debate, then the regulation being proposed is not stronger than the refrigerator’s. It is weaker, because the refrigerator at least has a test.

The second half of the question is harder. Who in the room has shipped a system that depended on such a test? The people who could write the AI equivalent of a refrigerant-leak check are evaluation engineers: the ones who already probe models for whether they will help synthesise a pathogen, whether they will deceive an overseer, or whether they will copy their own weights when given the chance. That work is unglamorous and, in places, already rigorous. It mostly sits inside the labs being regulated, because nobody outside offered those engineers a desk. A rulebook drafted in their absence will be shaped by the worst headline of the month rather than by the daily behaviour of the system.

A transmission note for the reader. When a regulatory proposal names an outcome, such as safety, control or the future, and does not name a measurement, it is asking for trust before it has earned the right to ask. The refrigerator earned that trust one boring failure mode at a time.

None of this is a reason to dismiss Hermer. The instinct is sound and the risk is real, and an office that holds lawyers to account is an odd place from which to say the law is behind. The people who drafted the fridge rules deserve the credit here, though. They had no grand theory. They had a lab bench and a technician who opened and closed the same door again and again, writing down the exact moment the seal gave way.