What arrives

Writing, images, music, teaching, advice, and conversation that a machine produced and no one can tell apart from a person’s. Some of it will be better. The economic loss is not that machines make good work; it is that people who would pay extra for human work lose the ability to know they are getting it, and stop paying.

The capability pages that drive it: persuasion and companionship and automating cognitive work.

What must exist first

Provenance and disclosure standards.

  1. A provenance chain for human work. A way to show that a person made something, which buyers can check and forgers cannot cheaply fake.
  2. Disclosure of machine work in anything sold, published, or used in a decision.
  3. Disclosure of machine conversation wherever it bears on money, health, law, belief, or a vote.
  4. Channels that pay for human work. Exchange, commission, office, and venture, each able to recognize work because a person did it (how contribution gets paid).
  5. Provenance over suppression. Machine work is labeled, not banned.

Why it is cheaper first

Demand for human contribution collapses once no one can tell. A provenance standard set while human work is still common is adopted by the people who already make it. Set after a decade of decline, there are fewer makers left to adopt it, and the market for human work has already priced itself as if it were gone.

Readiness states

State What it looks like
Started a provenance standard published; disclosure rules written for some fields
In place provenance marks in common use; disclosure routine in sales and public services
Tested demand and prices for verified human work holding while machine versions spread

What to start now

Adopt a provenance standard in public communication and public purchasing. Require disclosure of machine conversation in sales and services. Pay commission funds for work whose value lies in a person doing it: care, teaching, mentoring.

Where it is worked out