Operationalizing Human Sovereignty in High-Stakes AI Systems: The F-Request Protocol and Nf-Ledger as an Engineering Architecture for Oversight

Counts differ because each database indexes a different set of publications. We treat OpenAlex as the canonical count; Google Scholar is not shown (no API, and crawling it violates its ToS).

This paper proposes an operational software architecture grounded in the conceptual framework B+F=Nf, where B denotes functional processing (biological or algorithmic), F denotes the sovereign human witness, and Nf denotes the narrative trace produced when B passes through F. The paper introduces the F-Request Protocol as an engineering mechanism for human oversight in high- stakes AI systems, imposing a mandatory halt at semantically defined thresholds and requiring the user to provide a micro-narrative justification before endorsing a decision. The paper poses an exploratory research question: what is the preliminary effect size of mandatory justification and perceived accountability on hallucination acceptance rates in a simulated environment? The proposed pilot study is designed to estimate variance and the sample size required for a future randomized controlled trial; its results are not intended to support definitive conclusions about protocol efficacy. This study tests the behavioral impact of perceived accountability in a simulated setting and does not evaluate the technical or cryptographic infrastructure of the Nf-Ledger itself