AffectiveAI Research

Principles and limits

Responsible research begins by stating what it does not intend to do.

The programme privileges falsifiability, explicit uncertainty, human authority and a separation between research evidence and product incentives.

  1. Falsifiability. A hypothesis must be able to be tested, revised or rejected.
  2. Explicit uncertainty. An estimate is not the reality of a person.
  3. Human authority. Governance should preserve the human ability to decide, intervene and review.
  4. No social scoring. We do not build a ranking of people or infer private states as facts.
  5. Declared scope. Every public document should state what it does and does not establish.
  6. Separation of incentives. Research and product may share learning, not the same evidentiary threshold.