Hermer demands urgent law to control artificial intelligence
The Refrigerator Has Been Counted
Richard Hermer’s proposal benefits everyone now alive and everyone not yet born, by an amount no one can presently state. It burdens a small number of firms with the cost of being inspected. The arithmetic is uncomfortable because one column of the ledger is blank, and the blank is the whole scandal.
Let us count, beginning with the appliance. Mr Hermer observes that artificial intelligence is held to looser rules than a refrigerator. My colleagues laugh at this, as if he had said something whimsical. He has said something precise. The refrigerator is regulated not because it is dangerous, for it is among the most placid objects in the household, but because its dangers are countable. It may leak a gas, overheat a wire, or trap a child who climbs inside. Each pain has a known intensity, a known probability and a known population of kitchens. A legislator can weigh it in an afternoon. So the refrigerator has a standard for its door seal, a rating for its energy and an inspector with a clipboard. The law did not follow the danger. It followed the measurement.
That is my claim, and it is less comfortable than the one Mr Hermer makes. We have not neglected artificial intelligence because we think it harmless. We have neglected it because its harms are uncertain, remote and diffuse. In my own scheme of reckoning, those are exactly the qualities that make a pain easy to discount and hard to legislate against. A pain that is near and certain finds its statute quickly. A pain that is distant and doubtful waits in the anteroom with its hat in its hands while ministers attend to the refrigerator.
The strongest objection deserves its full weight, and it is a utilitarian objection, which is why I respect it. The benefits of these systems are near and fairly certain: diagnoses made sooner, drudgery lifted, knowledge put within reach of people who never had a library. The catastrophic harms Mr Hermer fears, the loss of control over the future itself, are speculative. A legislator who sacrifices a bird in the hand to guard against a dragon in the fog has done arithmetic badly. Heavy rules imposed now, in ignorance, may forbid the cure along with the poison, and the patients denied that cure will be real people with real pains.
I accept every premise and reject the conclusion. The objector assumes we are weighing a known benefit against an unknown harm. In fact we are weighing a benefit described by the people who sell it against a harm that no one is permitted to measure. The ledger exists, but it is kept in private by the firms whose interest lies in its contents not being read. Picture a refrigerator manufacturer who declined to say what gas circulated in the coils, who certified his own door seals, and who assured the public that any child found inside had climbed in of their own free will. We would not call that innovation. We would call it a defendant.
So the first law a rational legislator writes is not a prohibition. It is a law of counting: compulsory disclosure of what these systems can do, mandatory reporting when they fail, and inspectors with the same standing to open the cabinet that the appliance inspector already has. This costs the firms little. It costs the public nothing. And it converts the dimension of uncertainty, which currently excuses everyone, into something a calculus can use.
Here Mr Hermer’s insistence on international law stops being diplomatic ornament and becomes plain arithmetic. A faulty refrigerator harms one kitchen in one jurisdiction, and one jurisdiction can regulate it. The extent of a pain is the number of people who suffer it, and the extent of this one does not stop at Dover. Each person counts for one, whether they live in Leeds or Lagos. A national law that inspects only domestic machines while the same machines are built and run abroad is a door seal fitted to half a door. It also teaches every firm the oldest lesson in regulatory geography, which is to move the factory to wherever the inspector isn’t.
None of this requires us to believe in the dragon. It requires only that we stop letting the people who breed it decide whether it exists.
Somewhere tonight an inspector is kneeling on a showroom floor, pressing a strip of paper into the rubber gasket of a refrigerator door and pulling it out to test the grip. He is diligent, he is underpaid, and his work matters, because the paper tells him something true. In the building across the street, someone is releasing a system that may reshape how millions of people learn, work and decide, and no one is kneeling at its door with a strip of paper. They should be.