Independence policy.
Fide AI's credibility depends on keeping evaluation outcomes separate from donor interests, participant interests, related entities, and commercial incentives.
Policy posture
Research funding
Public benchmark research should be donor, grant, or public-interest funded wherever possible.
Participant-paid work
If participant-paid evaluations are offered, payment cannot affect methodology, scoring, publication, certification, appeals, or claims language.
Related participants
Any related participant requires enhanced disclosure, recusal, independent review, and non-conflicted signoff.
Public claims
Claims must preserve benchmark version, evaluated configuration, limitations, and the distinction between evidence and endorsement.
Compensation and access
Fide AI should build toward diversified nonprofit funding so public benchmark work is not dependent on the entities being evaluated. Scholarship and access support should help under-resourced institutions benefit from evaluation without turning standards into a pay-to-play market.
Non-interference rule
No donor, sponsor, participant, related entity, or paying customer may control benchmark scores, rankings, certification decisions, methodology freeze decisions, appeals, corrections, release timing, or publication signoff.