Information can remain available, portable, and historically accurate after the factual, legal, or institutional predicate that justified consequential reliance upon it has changed.
Research Poster · Governing Relationship State
Governing Relationship State: A Legal Framework for Human-Controlled AI Continuity
Micheal Charles Preble · Operant Dyad — Paper 4 · SSRN 7396659
Problem · Background · Research Question
Privacy law, agency law, consumer-reporting law, and records law already govern pieces of this problem; the paper synthesizes rather than replaces them.
What legal framework should govern continued use, transfer, restriction, revalidation, retirement, and reconstitution of persistent Relationship State?
Framework: the central object
The relevant legal object is not ownership of AI memory but lawful continuing reliance on persistent state.
Central diagram: eight governance duties
Visibility
Disclosure of the material state being relied upon, relevant source information, current purpose, and known restriction status.
Contestability
A meaningful way for the affected person to challenge the material basis of Consequential Reliance — broader than factual correction, narrower than a right to inspect the whole system.
Provenance
Access to material provenance sufficient to distinguish human declaration, imported record, observation, and machine inference.
Revalidation
Proportionate review of a Material Predicate when an actor with Legally Relevant Control Over Reliance knows, or should recognize, it may no longer be valid.
Restriction
Prospective recognition of legally effective revocation, expiration, or dispute, without automatically requiring deletion.
Retirement
Preservation of Covered Relationship State where lawful or required, while removing its ordinary eligibility to support current Consequential Reliance.
Transition review
Destination responsibility when state transfers: an express restriction, dispute, or material contextual change can trigger further review.
Material Reconstitution safeguards
Renewed review when substantially equivalent operative state is recreated from summaries, records, or derivative state after having been corrected, restricted, or retired.
Propositions / Observable Implications
- State ≠ consequence: a record can remain true as history while the consequence it once supported becomes unwarranted.
- Duties follow function, not custody: the actor with Legally Relevant Control Over Reliance bears the duty, not whoever first stored the state.
- Revalidation is event-triggered: elapsed time alone is weak evidence; a legally effective revocation, expiration, or failed verification is what should call continued reliance into question.
Evidence Required
A comparative legal synthesis, not new empirical data: GDPR Articles 18–20, the FCRA reinsertion regime (15 U.S.C. §1681i), 44 U.S.C. §3101 federal records law, HIPAA authorization rules (45 C.F.R. §164.508), the EU Data Act (Arts. 23–24, recital 86), and real case law including Van Patten v. Vertical Fitness Group (9th Cir. 2017) and Herbert Construction Co. v. Continental Insurance Co. (2d Cir. 1991).
Limitations
Institutional overlap between a broad framework and mature sector-specific systems is the most important unresolved policy question. Does not claim one institutional vehicle has been proven superior. Whether generalization across sectors is warranted depends on empirical prevalence not yet established.
Falsifiers / Open Questions
The strongest counterargument the paper engages directly: existing law may already be sufficient, in which case the proposal “adds nothing.” The paper's response is conditional — its case is strongest only at the residual coordination point where existing doctrines govern separate predicates but no clear mechanism keeps them synchronized across time, systems, or institutions.
Citation / QR
Suggested citation
Preble, Micheal Charles. “Governing Relationship State: A Legal Framework for Human-Controlled AI Continuity” Available at SSRN 7396659, 2026.
Preble, Micheal Charles. “Governing Relationship State: A Legal Framework for Human-Controlled AI Continuity” Available at SSRN 7396659, 2026.
research.perfinitive.com/publications/governing-relationship-state