Why “AI against regulation” loses
The common picture puts AI on one side and regulation on the other, and asks whether the law can keep up. It cannot. Laws take years to draft, pass, and enforce; capabilities can spread in months. The people a law is meant to protect have usually lost something before it takes effect. And a race framed this way invites two bad answers: give up because the law will never catch up, or try to stop the technology outright, which no single country can do.
The framing is also too narrow. Regulation is one kind of readiness among several, and often not the most important.
What readiness includes
| Kind of readiness | What it means | Example |
|---|---|---|
| Law and rules | what is permitted, required, or barred | a right to contest any automated decision that affects one’s standing |
| Institutions | bodies that can act, appeal, measure, and enforce | independent appeal bodies; the estimator pool |
| Tools people use | means for participants to see, contest, and leave | a plain account of what you receive; portable data; exit routes |
| Developed human capacity | skills and judgment people still hold | apprenticeship pipelines; competence reserves for critical systems |
| Preserved alternatives | other ways of doing essential things | a second provider; manual fallbacks; forkable systems |
| Material security | what keeps people from trading jurisdiction for survival | the floor, paid before displacement peaks |
A society can be ready even where its laws lag, if its institutions, tools, people, and alternatives are in place. It can be unready with excellent laws, if no one can use them, no body can enforce them, and the only way of doing something essential runs through one system.
Why before is cheaper than after
Almost every kind of readiness costs less to build before a capability becomes ordinary. Three reasons:
- Dependence locks in. Once a whole society routes an essential function through one system, building an alternative means competing with something everyone already uses, financed by the very provider it would replace. Before that, a second provider is an ordinary contract.
- Skills decay. Human competence that is not used fades within years, and the people who could teach it retire. An apprenticeship pipeline kept running costs little; one restarted after a generation without it may be impossible.
- Leverage fades. Today AI firms depend on public inputs: grid connections, permits, public data, spectrum. Conditions attached at those inputs cost the public nothing. As firms build their own, that leverage decays, and the same conditions later would have to be won against stronger opposition.
There is a fourth reason, less visible: people lose the standing to ask. A population already displaced, dependent, or desperate is in a weak position to demand anything. Readiness built while people still hold work, competence, and votes is readiness they can shape. Readiness built after is readiness done to them.
Readiness and pace
Commensurate pace says capability should arrive no faster than an ecology can absorb it. Readiness says what absorbing it requires. The two work together: where readiness is in place, a capability can spread quickly and safely; where it is not, pace is the only protection left, and the time it buys is only worth something if it is used to build what is missing. See when capability outruns readiness.
Pace buys time. Readiness is what the time is for.