AI Governance research, frameworks, and working papers by Dr. Tuboise Floyd, PhD — Founder of Human Signal, Editor in Chief of The AI Governance Record, and Host of The AI Governance Briefing.
This repository is the canonical research archive for the Human Signal framework family: GASP™, TAIMScore™, Failure Files™, the Workflow Thesis, the Trust Gap, the L.E.A.C. Protocol™, PSA®, AIaPI™, Hyperprompt™, and Noise Discipline.
Thesis. Most institutions will not fail because of a bad AI model. They will fail because of a broken governance structure around it.
Brand. Independence is not a feature. It is the product.
Dr. Tuboise Floyd, PhD — Founder and Chief Sensemaking Officer, Human Signal. PhD in Adult Education / Systems Theory (Auburn University, 2010). TAIMScore™ Certified Assessor (HISPI, March 2026). Member, HISPI Advocacy & Education Working Group (Project Cerebellum AI Think Tank).
- Website: humansignal.io
- Mission page: humansignal.io/mission_dr_tuboise_floyd
- ORCID iD: 0009-0008-0055-072X
- SSRN author page: ssrn.com/author=11021388
- U.S. Copyright Office Registration (PSA®): TXu002503385
- Podcast: The AI Governance Briefing
- Newsletter: The AI Governance Record
Dr. Floyd is the founding pedagogical theorist of AI governance. His doctoral work in andragogy and self-directed learning theory is the academic engine behind every framework in this repository.
Each paper below is a working paper hosted in this repository. Where an SSRN preprint exists, the SSRN identifier is listed alongside the local PDF. Cite by author and year; full citation blocks are provided per paper.
File: Dr_Tuboise_Floyd_Workflow_Thesis.pdf
Companion: Dr_Tuboise_Floyd_Workflow_Thesis_Position_Paper.pdf
Framework page: humansignal.io/frameworks/workflow-thesis
Abstract. Institutions deploying AI fail not because of underperforming models but because of broken governance structures around those models. The Workflow Thesis reframes the institutional AI problem from a technology procurement question to a structural accountability question: who owns the decision, what is the escalation path, and what accountability exists without a vendor in the room. This paper establishes the thesis, defines its diagnostic surface, and connects it to the GASP™ instrument and the Trust Gap typology.
Cite as:
Floyd, T. (2026). The Workflow Thesis: Institutions Fail at Governance Structure, Not Model Performance. SSRN Working Paper. https://ssrn.com/abstract=6644860. DOI: 10.2139/ssrn.6644860
File: Dr_Tuboise_Floyd_GASP.pdf
Framework page: humansignal.io/frameworks/gasp
Trademark status: GASP™ (registered).
Abstract. GASP™ (Governance As a Structural Problem) is a paid diagnostic instrument for institutions deploying AI. Rather than auditing models, GASP™ audits the structure around them: decision ownership, escalation paths, vendor-independent accountability, and the workflow conditions that determine whether an AI system can be governed at all. GASP™ is the operational expression of the Workflow Thesis and the entry point for the Human Signal advisory engagement model.
Cite as:
Floyd, T. (2026). GASP™: Governance As a Structural Problem. Human Signal Working Paper.
File: Dr_Tuboise_Floyd_Trust_Gap.pdf
Framework page: humansignal.io/frameworks/trust-gap
Abstract. The Trust Gap is a two-level typology for institutional AI accountability failures. Level 1 — Structural Absence: no governance structure exists where one is required. Level 2 — Structural Insufficiency: a structure exists but is not admissible against the decisions the institution is making. The operative principle: permitted is not the same as admissible. The Trust Gap supplies the diagnostic vocabulary that GASP™ assessments resolve into.
Cite as:
Floyd, T. (2026). The Trust Gap: Structural Absence and Structural Insufficiency in Institutional AI Governance. Human Signal Working Paper.
File: Dr_Tuboise_Floyd_LEAC_Protocol.pdf
Framework page: humansignal.io/frameworks/leac
Abstract. The L.E.A.C. Protocol™ identifies four physical-infrastructure constraints that bound every AI deployment regardless of model choice: Lithography, Energy, Arbitrage, and Cooling. Governance frameworks that ignore the physical substrate produce policy artifacts that cannot be enforced. L.E.A.C. is the substrate-aware companion to GASP™ and the Workflow Thesis.
Cite as:
Floyd, T. (2026). The L.E.A.C. Protocol™: Lithography, Energy, Arbitrage, Cooling. Human Signal Working Paper.
File: Dr_Tuboise_Floyd_PSA_AIaPI_Framework.pdf
U.S. Copyright Office Registration: TXu002503385
Abstract. PSA® (Presence Signaling Architecture) is a federally registered framework for adaptive signal discipline: Signal → Content → Calibration → Adaptation. AIaPI™ (AI as a Pedagogical Instrument) is the educational application layer. Together they define how an institution converts AI exposure into structured learning rather than passive consumption. PSA® is the foundational IP under the Visible Human research house.
Cite as:
Floyd, T. (2026). PSA®: Presence Signaling Architecture, and AIaPI™: AI as a Pedagogical Instrument. U.S. Copyright Office Registration TXu002503385.
Files:
Dr_Tuboise_Floyd_The_Pedagogy_Problem_in_AI_Governance_SSRN.pdfThe_Pedagogy_Problem_in_AI_Governance_Position_Paper.pdf
Position paper (web): humansignal.io/position-paper
Abstract. AI governance is currently treated as a compliance problem. It is, in fact, a pedagogy problem: institutions cannot govern what their decision-makers cannot interpret, and they cannot interpret what they have not been taught using methods grounded in adult learning theory. This paper argues that governance literacy is the missing prerequisite for every existing framework — NIST AI RMF, EU AI Act, and TAIM included — and proposes andragogy-grounded pedagogical scaffolding as the resolution path.
Cite as:
Floyd, T. (2026). The Pedagogy Problem in AI Governance. Human Signal Working Paper.
File: Dr_Tuboise_Floyd_Hyperprompt_Signal.pdf
Coinage attribution: Term Hyperprompt™ coined by Steven Cash; framework development by Dr. Floyd.
Abstract. Hyperprompt™ is the signal-discipline instrument inside the PSA® family. Where conventional prompt engineering optimizes a single exchange, Hyperprompt™ governs the architecture of repeated exchanges — calibrating signal density, adaptation cadence, and presence under load.
Cite as:
Floyd, T. (2026). Hyperprompt™ Signal. Human Signal Working Paper.
File: Dr_Tubosie_Floyd_Noise_Discipline.pdf
Abstract. Noise Discipline is the governance counterpart to signal architecture: the operational practice of refusing inputs that degrade institutional decision quality, regardless of source authority. It supplies the editorial standard behind The AI Governance Record and the institutional standard behind GASP™ engagements.
Cite as:
Floyd, T. (2026). Noise Discipline. Human Signal Working Paper.
THESIS LAYER
└── The Workflow Thesis ──────────── institutions fail at structure, not model
DIAGNOSTIC LAYER
├── GASP™ (Governance As a Structural Problem) ── paid instrument
├── The Trust Gap ──────────────────────────────── absence vs. insufficiency
└── L.E.A.C. Protocol™ ─────────────────────────── physical substrate audit
PEDAGOGICAL LAYER (Visible Human house)
├── PSA® (Presence Signaling Architecture) ─────── USCO TXu002503385
├── AIaPI™ (AI as a Pedagogical Instrument)
├── Hyperprompt™ Signal
└── Noise Discipline
POSITION LAYER
└── The Pedagogy Problem in AI Governance
All papers in this repository are working papers by Dr. Tuboise Floyd, PhD. Default citation form:
Floyd, T. (2026). [Paper Title]. Human Signal Working Paper. https://github.com/drtfloyd/Research
Where an SSRN preprint exists for the paper, prefer the SSRN identifier:
Floyd, T. (2026). [Paper Title]. SSRN Working Paper. https://ssrn.com/abstract=[ID]. DOI: 10.2139/ssrn.[ID]
- GASP™ is a registered trademark of Dr. Tuboise Floyd / Human Signal.
- PSA® is a federally registered work, U.S. Copyright Office TXu002503385.
- TAIMScore™, Failure Files™, L.E.A.C. Protocol™, AIaPI™, Hyperprompt™ are claimed marks under active use.
- Human Signal™ is a claimed mark under active use.
The frameworks in this repository are the intellectual property of Dr. Tuboise Floyd. Working papers are posted for academic citation, peer engagement, and discoverability. Commercial use, derivative training, or licensed adaptation requires written agreement.
- Web: humansignal.io
- Newsletter: theaigovernancerecord.com
- Podcast: theaigovernancebriefing.com
- LinkedIn: linkedin.com/in/theaigovernancebriefing
Independence is not a feature. It is the product.