Hermer demands urgent law to control artificial intelligence

The crisis arrived on a timeline that assumed institutions capable of processing it at a speed they last achieved in the summer of 1945, when a handful of physicists and a smaller handful of statesmen found they had perhaps a few years to invent a whole grammar of restraint. The institutions, naturally, processed it at the speed they are equipped for, which is the speed of the household appliance.

That is the instructive part of Richard Hermer’s complaint that artificial intelligence answers to fewer rules than a refrigerator. He offers it as a scandal. It is better read as a measurement. Legislators did not regulate the refrigerator because they understood cold. They regulated it because it had several generations in which to misbehave in plain sight. Early machines leaked toxic gases into kitchens. The safer chemistry that replaced them was hailed as a triumph and then, decades on, found to be thinning the sky. More than half a century separated the invention of those refrigerants from the Montreal Protocol that curtailed them. The fridge in a modern showroom, with its energy label and certified compressor, is a monument to accumulated injury, patiently litigated, and not to foresight.

So the comparison flatters the refrigerator. It mistakes the end of a long education for its beginning.

Mr. Hermer is a lawyer and a law officer of the Crown, and he reaches for the remedy his training supplies: urgent public debate and robust international law. One should not sneer at this, because the strongest case for it is the nuclear case. Treaties, inspectors and a dense lattice of agreements did, imperfectly, keep the most dangerous dynamo of the last century from being used in anger after its first two demonstrations. If law caught up with the atom, why not with the algorithm?

The answer is that the atom was obliging enough to be heavy. It required enrichment halls visible from orbit, ore dug from particular mountains, and a population of specialists who could be counted. International law works by counting things at borders: warheads, centrifuges, tonnes. The new force has no tonnage. Its weights are a file, and a file can be copied in the time it takes a delegation to agree on the seating plan, with the copy as capable as the original. The treaty instrument was calibrated for a force that moved at the speed of freight. The present one moves at the speed of a download, and nobody has yet proposed customs inspection for arithmetic.

Here the entropy becomes measurable, if not in numbers then in direction. Each year the capability compounds. Each year the institutional response adds a consultation, a framework, a voluntary commitment signed in a hotel ballroom, and an expert panel instructed to report the following spring. These are not failures of will. They are the procedures that worked last time, and every institution meets a new force by first deploying its full catalogue of answers to the old one.

There is a quiet comedy here that Mr. Hermer, being serious, cannot permit himself. The public debate he calls for will be conducted largely on platforms ranked, summarised and increasingly drafted by the very systems under discussion. The defendant has been asked to take the minutes. One does not fault him for the request, since there is no other room in which to hold the meeting.

What he has right, and what deserves more attention than the refrigerator, is his warning that we could lose control of the future. A man of law speaks of control because law presumes a sovereign capable of comprehending what it governs. The historian’s question is narrower: whether comprehension is still accelerating at all, or whether we have entered the first period in which the governed object learns faster than the governor reads. If it is the latter, then urgent debate is not wrong, only late in the way all debates are late. Robust law will arrive as the fridge label arrived, describing with great precision a machine that has already left the showroom.

A man of my generation learned that his education expired roughly every twenty years, and the interval appears to have shortened. Somewhere this week a junior official is drafting a definition of “artificial intelligence” for a treaty annex. By the time the second committee has reviewed the ink, the thing defined will have shipped a new version, and the definition will be filed, cold and orderly, beside the regulations for the refrigerator.