Case study

Study provenance: Run with an AskSociety digital population. No human validation layer was included. Findings are directional decision evidence, not observed human behavior.

What happened when 576 residents were asked whether AI is good for humanity

Study typeAttitude / Society
Population576 residents
MethodBlind society study
576independent residents
72psychological foundations
1society with disagreement left intact
Introduction

A question about AI sounds like something a model should answer cleanly. A society should not.

The challenge

A polished consensus can erase the very disagreement the question is trying to understand.

Questions about technology and humanity carry different fears, incentives, experiences, and definitions of progress. A single synthesized answer tends to smooth those differences into something reasonable and balanced. That is precisely what this study was designed not to do.

The study

Ask one society and leave the disagreement intact.

The question was put to 576 independent residents built on 72 psychological foundations. The population was allowed to disagree rather than being forced toward a consensus, preserving the competing reasons behind support, resistance, ambivalence, and uncertainty.

The results

The result was a map of disagreement rather than a single opinion.

  • 576 residents responded independently.
  • The population was grounded in 72 psychological foundations.
  • Disagreement remained visible instead of being collapsed into one synthetic consensus.
  • The output showed why a society can be more useful than one polished model answer for contested questions.
The read
“The point was not to make the society agree. It was to understand the structure of why it did not.”
The read
Your turn

What decision are you sitting on?

Bring a decision →