Portable memory and context infrastructure
Persistence, interoperability, provenance, preference management, portability, and user control. Perfinitive concentrates further downstream on present authority and legitimate continued use.
Related research
The surrounding field is already moving.
This page maps the scholarship, legal authorities, standards, and neighboring programs that inform, precede, or pressure the Perfinitive research program.
Authorship boundary: Every record on this page is an external source produced by other researchers, courts, legislatures, standards bodies, or organizations. None is Micheal Preble’s research. “Referenced in” identifies relevance to a Perfinitive paper or research direction; it does not imply authorship, affiliation, priority, or endorsement.
Research map
Where Perfinitive sits
The map makes foundations and close neighbors visible before stating the narrower questions Perfinitive is developing.
Extended cognition
External systems can genuinely extend human capability. Perfinitive asks when extended performance warrants a human-advancement claim.
Human–AI teaming and complementarity
Joint performance and coordination. Perfinitive separates configuration performance from the trajectory of the human participant.
Meaningful human control and autonomy
Practical authority, responsibility, contestability, and intervention provide evidence constructs for governed extension.
Automation, delegation, and skill change
Short-run gains can coexist with loss of practice, retained-capacity decline, and path dependence.
Capability approach and human development
Resources are not identical to substantive freedoms. Perfinitive applies that separation to AI-progress claims.
Digital inequality and institutional power
Average benefit and formal access can conceal unequal capability, burdens, opacity, and power.
Longitudinal relational interaction
Repeated interaction, memory, adaptation, and relationship effects provide empirical neighbors for relationship-state research.
Taxonomy
Three groups the map is organized around
Every external source below sits in one of three groups, tagged by its closest overlap with the program.
cognition · agency · teaming · capability
memory · context · longitudinal interaction · portability
control · authority · inequality · institutions · economics
Institutional signals
Where institutions are pointing
Reports, dialogues, and standards work from multilateral and standard-setting bodies now naming the same territory: persistent AI interaction, human oversight, and governance of accumulated state.
United Nations
AI Advisory Body reporting on global governance gaps for AI systems.
UNESCO
Recommendation on the Ethics of Artificial Intelligence and human oversight guidance.
ILO
Labor-facing work on AI, skill formation, and the future of work.
WIPO
IP-facing conversation on AI-assisted authorship, inventorship, and provenance.
Global Dialogue on AI Governance
Emerging multilateral forum for coordinating AI governance questions across jurisdictions.
Institutional signals are evidence about where important questions are moving—not evidence that Perfinitive’s answers are correct.
Public lineage
This program begins inside an established tradition.
Its lineage runs from human–computer symbiosis and augmentation through distributed cognition, automation and human factors, relational agents, long-term HRI, human–AI teaming, co-adaptation, distributed agency, and contemporary evidence on performance and learning.
LickliderEngelbartDistributed and extended cognitionAutomation and human factorsRelational agents and long-term HRIHuman–AI teaming and co-adaptationGenAI performance and learning evidencePerfinitive’s longitudinal governance questions
Claim boundary: Perfinitive does not claim as new the broad ideas of long-term human–AI relationship, co-adaptation, shared or distributed agency, persistent memory, memory portability, meaningful human control, or the distinction between assisted performance and human capacity.
Research credit matrix v0.1
Credit first. Residual question second.
The four labels describe each source’s role in this program. They are research-mapping terms, not judgments of quality or priority.
Foundational
Supplies a premise or tradition on which the program depends.
Direct predecessorEstablishes a close component of the problem before Perfinitive.
Empirical neighborProvides evidence that constrains or tests a nearby claim.
Current close neighborDevelops a substantially overlapping contemporary question.
J. C. R. Licklider · Man-Computer Symbiosis (1960)
- What they established
- A durable augmentation tradition centered on complementary human and machine capacities.
- What we use from it
- The premise that the configured relationship can be a meaningful unit of analysis.
- What we do not claim
- Perfinitive did not originate human–computer symbiosis or augmentation.
- What question remains
- How should accumulated relationship state remain governable across time and institutional change?
Douglas Engelbart · Augmenting Human Intellect (1962)
- What they established
- Human capability can be reorganized through tools, language, methods, and training as a system.
- What we use from it
- Augmentation is more than a faster output; it concerns the capability-bearing configuration.
- What we do not claim
- Perfinitive did not originate human-intellect augmentation.
- What question remains
- Which part of an observed gain became durable human capability, and who controls the remainder?
Bickmore & Picard · Long-Term Human-Computer Relationships (2005)
- What they established
- Conversational systems can deliberately form and maintain relational continuity over repeated interaction.
- What we use from it
- Repeated interaction and remembered history can change the later relationship.
- What we do not claim
- Perfinitive did not originate long-term relational agents.
- What question remains
- What authority governs the later use, correction, transfer, and retirement of accumulated relationship state?
Parasuraman & Riley · Humans and Automation (1997)
- What they established
- Automation can be used, misused, disused, or abused, with consequences for human performance and control.
- What we use from it
- System performance cannot be interpreted apart from the human role the system reshapes.
- What we do not claim
- Perfinitive did not originate automation-complacency or out-of-the-loop concerns.
- What question remains
- What evidence shows whether repeated AI use preserves or converts human capacity over time?
Santoni de Sio & van den Hoven · Meaningful Human Control (2018)
- What they established
- Human control depends on more than nominal oversight; systems must track relevant human reasons and support responsibility.
- What we use from it
- Practical authority and responsibility are stronger tests than interface-level control.
- What we do not claim
- Perfinitive did not originate meaningful human control.
- What question remains
- How should present authority be reevaluated when relationship state moves across roles, systems, or purposes?
Noy & Zhang · Productivity Effects of Generative AI (2023)
- What they established
- Generative AI can improve speed and quality on bounded writing tasks.
- What we use from it
- Assisted performance gains are real and measurable.
- What we do not claim
- A measured task gain alone establishes durable human advancement.
- What question remains
- What persists in the person when assistance is removed or the configuration changes?
Passalacqua et al. · Practice With Less AI Makes Perfect (2024)
- What they established
- The degree of automation during training can alter motivation, engagement, and skill acquisition.
- What we use from it
- Short-run assistance and longer-run human formation must be measured separately.
- What we do not claim
- One study settles the effect across tasks, systems, or populations.
- What question remains
- Which longitudinal measures reveal durable conversion, dependency, or recovery?
Yun, Taranova & Wang · Does My Chatbot Have an Agenda? (2026)
- What they established
- A month-long study in which conversational control emerged as a shared, turn-by-turn experience.
- What we use from it
- Agency in sustained interaction may be co-constructed rather than located in one actor alone.
- What we do not claim
- Perfinitive did not originate shared or distributed agency.
- What question remains
- How do authority, correction, provenance, and exit operate within a persistent distributed configuration?
Published position
What remains after predecessor credit is given
These are bounded positions in the published or submitted corpus. They are invitations to testing, not declarations that the surrounding components are new.
Emergent NoveltyDyad-specific attribution using NAT and ENR; preliminary single-case evidence.
Operant DyadThe specific task-bound human–AI working unit assembled from repeated goals, correction, retained state, inspection, and revision.
The Persistent PrimitiveA person-centered canonical state, authorized views, and transport governed by provenance, correction, revocation, and purpose.
Authority Does Not TravelDestination-side reevaluation of legal operability when relationship state crosses a technical or institutional boundary.
Governing Relationship StateMaterial reconstitution: what continuing reliance requires after authority, restriction, or correction changes.
Beyond PerformanceA claim-specific diagnostic method for testing human-advancement claims; not the elementary distinction between assistance and capacity.
LiO v0.1Deterministic composition of supplied permissions across indispensable bridges; not representation, adjudication, or set-theoretic novelty.
External initiative · close research neighbor
Portable, user-governed memory
Stanford’s Portable Memory initiative addresses persistent personal context, portability, provenance, user control, and movement across systems. That work constrains any broad Perfinitive claim about portable or user-controlled AI memory.
Perfinitive’s narrower continuity question begins after successful transfer: whether a receiving system has an adequate present basis to use the state under changed provider, role, purpose, time, institutional, or jurisdictional conditions.
Stanford Digital Economy Lab: Portable Memory →External initiative · economic measurement neighbor
AI Economic Indicators
Stanford’s AI Economic Indicators assembles regularly updated measures of employment and wages, aggregate economic outcomes, consumer surplus, hiring and skills, and AI usage. It provides the economic observation layer that the proposed Perfinitive work should complement rather than reproduce.
Perfinitive’s proposed addition begins beneath a measured performance gain: whether the gain became durable human capability, remained system-supplied, depended on governed relationship state, or changed how economic value and risk were distributed.
Stanford Digital Economy Lab: AI Economic Indicators →External research library
A–Z sources, authorities, and standards
Search the full provenance record or filter it by category and the Perfinitive work in which it is referenced.
Loading external records…
Perfinitive research boundary
What remains distinctive enough to test
This library is a provenance record, not a novelty register. Labels such as foundational, adjacent, overlapping, or direct pressure describe how an external source bears on a Perfinitive claim. They do not claim authorship, ownership, priority, patentability, agreement, or legal conclusion.
Across the program, two narrower contributions remain open for testing. The continuity work asks what authority-relevant information must accompany relationship state and which changes trigger reevaluation. The economic work asks how much assisted performance becomes durable human capability, how much remains within a governed relationship, and how the resulting value and risk are distributed.