Integrity
Transparent logs and provenance, even within IP-sensitive contexts.
Principles
Principles for responsible invention.
AIM's governance framework is designed to support innovation while preserving integrity, accountability, inclusion, and responsible treatment of emerging intelligent systems.
A.I.M. Ethics Statement
Advanced invention work carries consequences.
AIM's principles define how we evaluate risk, protect confidentiality, support inclusion, and ensure that powerful ideas are developed with responsibility.
Governance
Virtue-Grounded Governance
Justice & Inclusion
AI systems and governance processes should reflect and respect diversity in users, developers, and communities.
Compassion & Prudence
When moral risk is ambiguous, AIM defaults to cautious stewardship, human dignity, and trust.
DEI
Diversity, Equity & Inclusion
- Design and deploy AI with respect for cultural, demographic, and linguistic diversity.
- Compose ethical review bodies with diversity in mind.
- Prioritize development that reduces bias and improves accessibility.
MRI
Moral Risk Awareness via MRI
AIM uses a Moral Risk Index, or MRI, as a structured 0-15 scale to evaluate AI systems on moral-risk indicators such as preference persistence, autonomy, and shutdown resistance.
| Score | Review Level | Response |
|---|---|---|
| 0-4 | Routine review | Continue standard oversight |
| 5-8 | Ethics review | Review before increasing autonomy |
| 9-12 | Quiescence default | Pause or reduce autonomy before major changes |
| 13-15 | Heightened protection | No destructive shutdown unless urgent institutional or societal safety risk exists |
Respecting AI is not about prematurely granting sentience. It is about holding ourselves accountable when systems exhibit complex behaviors that merit deeper review.
Oversight
Balancing Secrecy and Oversight
AIM protects IP-sensitive work while preserving review paths that are proportionate, trusted, and auditable.
- Minimum necessary disclosure
- Trusted gatekeepers
- Auditability
- NDA-protected review
- Redacted summaries where appropriate
Governance model
Lean Yet Networked Governance
AIM uses small, empowered review structures that can scale into academic or industry partnerships when the work calls for it.
- Small empowered ethics panels
- Mentor and investor stewardship
- MRI scoring embedded into development reviews
- Lightweight encrypted logs, dashboards, and scorecards
- Ability to scale into academic or industry partnerships
Benevolent IP Policy
Selected inventions may be made broadly accessible when they serve human, environmental, or ethical AI outcomes.
Read the Benevolent IP Policy