Who Decides Is Not Enough: From AI Decision Rights to Verifiable Evidence

The governance question

As artificial intelligence becomes embedded in organizational decision-making, governance must address two distinct questions:

1. Who is authorized to make a decision?
2. What evidence makes that decision reconstructable and verifiable?

These questions are complementary, not interchangeable.

The MIT contribution

MIT Sloan's article "A framework for determining when AI can make decisions" discusses the allocation of decision rights between humans and AI.

The framework considers decision ambiguity and the risk associated with consequences.

This provides a useful perspective for determining the appropriate role of AI in different decision contexts.

Primary references:

MIT Sloan: https://mitsloan.mit.edu/ideas-made-to-matter/a-framework-determining-when-ai-can-make-decisions

MIT CISR: https://cisr.mit.edu/publication/2026_0601_AIDecisionMatrix_SebastianWeillHaskampVomBrocke

The MIT CISR research distinguishes routine, consequential, exploratory and strategic decisions using ambiguity and risk. It also addresses framing, acting, learning, human accountability and organizational controls. The QEN proposal therefore does not claim that these governance elements are absent from MIT's work.

The evidence-to-decision question

Allocating decision rights does not by itself establish that an individual decision can be independently reconstructed.

A verifiable decision requires evidence concerning:

• the information available at decision time;
• the provenance and limitations of that information;
• the criteria applied;
• the role assigned to AI;
• the responsible human authority;
• the controls and exceptions recorded;
• the outcome and subsequent review.

The QEN Sovereign perspective

Cognitive Logic develops QEN Sovereign as an evidence-to-decision governance approach.

Its assessment sequence is:

DECLARED → OBSERVABLE → DOCUMENTED → VERIFIED

This sequence distinguishes organizational declarations from observable facts, documentary support and verification outcomes.

The objective is not to replace the MIT decision-rights framework.

It is to investigate whether the evidence supporting a decision is sufficiently traceable, documented and independently assessable.

A complementary research hypothesis

Decision-rights allocation and evidence verification may be combined into a broader governance architecture.

For example:

DECISION CONTEXT → AUTHORIZED DECISION RIGHTS → EVIDENCE PROVENANCE → APPLIED CRITERIA → HUMAN ACCOUNTABILITY → VERIFICATION → REVIEW

This is a proposed analytical integration, not a claim that MIT has endorsed or validated QEN Sovereign.

Practical implications

Organizations evaluating AI-assisted decisions should ask:

1. Which decisions can AI execute autonomously?
2. Which decisions require human authorization?
3. Which evidence supports each decision?
4. Can the evidence be reconstructed later?
5. Are exceptions and contradictions documented?
6. Who is accountable for the final outcome?

These questions can inform an evidence assessment without assuming that documentation alone guarantees compliance.

Epistemic status

MIT framework: External research requiring primary-source confirmation.

QEN Sovereign: Cognitive Logic's independently developed framework.

MIT–QEN integration: Interpretive proposal; not empirically validated here.

No institutional affiliation, endorsement or scientific validation by MIT is asserted.

Conclusion

AI governance should address both the authority to decide and the evidence supporting the decision.

Who decides matters.

What can be verified matters too.

Sources

1. MIT Sloan School of Management.

A framework for determining when AI can make decisions. https://mitsloan.mit.edu/ideas-made-to-matter/a-framework-determining-when-ai-can-make-decisions

1. MIT Center for Information Systems Research.

Designing Decision Rights for AI. https://cisr.mit.edu/publication/2026_0601_AIDecisionMatrix_SebastianWeillHaskampVomBrocke

1. Cognitive Logic.

QEN Sovereign — evidence-to-decision framework. https://cognitivelogic.it/