The S · E · E · N Research Cycle
A four-phase applied research protocol built for community-based study of responsible AI — designed for depth, honesty, and return of knowledge.
See
Sit with the environment before naming the question.
Every study begins in place. We spend time with the organization, its people, and the communities it serves — not to gather data, but to understand what a good question would even look like. Nothing is instrumented until the environment is understood.
- Site immersion with the community partner
- Stakeholder mapping and consent conversations
- Prior-work and policy landscape review
Examine
Co-write the research question with the people who live inside it.
Research questions are drafted together with the partner and reviewed by our advisory network. Methods are chosen for the context — mixed, qualitative, quantitative, ethnographic — not for methodological fashion. The protocol is written to be inspected.
- Community-co-authored research protocol
- Methods matched to context, not to trend
- Ethics and data-stewardship plan in writing
Experiment
Study AI as it is actually used.
We study systems in place — pilots, deployments, and everyday use — with instrumentation the community can inspect. Findings are held provisionally and revisited in the field. We are willing to be wrong slowly rather than right quickly.
- Field-based data collection with partner staff
- Iterative interpretation with community reviewers
- Auditable instrumentation and analysis logs
Navigate
Return the knowledge before it travels.
Findings are returned to the community partner first, in a form they can use — briefings, plain-language reports, and internal working sessions — before any external publication. Only after the return of knowledge does the study enter the scholarly and public record.
- Return-of-knowledge briefings with the partner
- Community-facing report and working session
- Working paper and public archive publication
What we hold ourselves to.
Sponsors fund studies. They do not shape, edit, or embargo findings.
Community partners receive findings first, in a form they can use, before public release.
Protocols, methods, and known limitations are published alongside every study.
A minimum one-year study horizon. Four partners per year, no more.
"The measure of an applied research institution is not what it publishes. It is whether the people who lived inside the study would let us come back."
— Institutional principle