The Simulated Public
We are beginning to ask people who do not exist
SIGNAL · AI / MARKET / CULTURE
For decades, market research followed a familiar sequence.
A company had a question.
It found people.
It asked them.
It tried to infer what a larger public might do.
That sequence is beginning to bend.
Synthetic respondents — AI-generated personas designed to approximate the behavior or opinions of particular groups — are moving from experimental curiosity into practical market-research workflows.
Companies can now test a product idea, advertising concept, price or interface against simulated respondents before recruiting real participants.
The attraction is obvious.
They are faster.
They are cheaper.
They can be multiplied almost without limit.
And they are always available.
But the interesting signal is not that market research is becoming automated.
It is that the public can now be consulted before the public arrives.
Structural Shift
Research is moving from observing real populations toward rehearsing their possible responses in synthetic populations.
This is a subtle but important change.
Traditional personas were descriptions.
Synthetic personas respond.
Put enough of them together and something new appears:
not a person,
but a simulated public.
The first-order benefit
Synthetic publics can be useful.
They can help teams explore assumptions before spending money on human research.
They can expose obvious weaknesses.
They can test multiple scenarios quickly.
They can make early-stage experimentation much cheaper.
Used carefully, the synthetic public could become a rehearsal space.
That may be valuable.
But the second-order effect is more interesting.
When rehearsal becomes selection
Imagine that a company tests two products.
The simulated public strongly prefers Product A.
Product B scores poorly.
The company therefore launches only A.
Later, real consumers buy A.
The simulation appears to have been correct.
But something is missing.
Real consumers never encountered B.
So did the synthetic public predict reality?
Or did it participate in deciding which reality became available?
This creates a loop:
Simulation
→ Prediction
→ Decision
→ Real-world availability
→ Human behavior
→ New data
→ Simulation
At that point, the simulated public is no longer only measuring the world.
It is entering the machinery that produces it.
What to observe next
Watch for synthetic populations moving beyond market research into:
- product development
- advertising
- entertainment
- public policy
- healthcare
- workforce planning
- community design
- political messaging
The key question is not whether synthetic respondents become perfectly accurate.
It is whether institutions begin trusting them enough to make consequential decisions before real people have had the opportunity to respond.
That threshold matters more.
Connected
The Synthetic Person
One artificial social actor.
The Simulated Public
A population composed of artificial actors.
The Pre-Selected Future
What happens when predictions from those populations begin determining which possibilities reach reality.
Signal to watch
The decisive moment will not be when a synthetic public can imitate us perfectly.
It will be when an institution says:
We already know what people will want.
And acts before asking them.
Continue → The Pre-Selected Future
If simulated reactions begin determining products, investments, jobs, research and culture, prediction stops being merely descriptive.
It becomes part of allocation.
