Overview
As AI-assisted claims become pervasive across research and industry in 2026, a growing epistemic gap has emerged: such claims can appear authoritative when evidence, analysis, human authorization, presentation, and correction history refer to different states. Provenance, attestation, and transparency mechanisms expose history but, on their own, do not specify the publication transition examined here.
This work introduces Publication Authority โ defined as an exact-state, non-transferable, single-use publication capability โ and instantiates it in PAC-2026 (Publication-Accountability Calculus), a machine-readable AIJIM Protocol candidate.
Core Contributions
The authors evaluate PAC-2026's fourth bounded semantic freeze (SF-4), a fixed-profile specification designed for replaceable bindings. Six obligations govern the framework:
- Evidence
- Runs and artifacts
- Measurement disclosure
- Authorization
- Surface correspondence
- Lifecycle continuity
- Ten models explored 110,764 safe reachable states.
- 76 unsafe configurations produced the expected violation or observer countermodel.
- A reader surface passing its correspondence check cannot authorize publication unless the accepted record admits that surface.
- SF-4 separates evidence horizon from verification time and rejects an authentic but causally invalid authorization.
- A historical predecessor path reproduced 17 frozen authorization-successor outcomes.
- A later in-house, instance-blind test of known case classes matched all 183 scored expectations.
- Same-host package execution reproduced its 240 archived observations.
- Authors: Torsten Olivi Tiltack, Yifei Dong, Kun Yu, Xu Wang, Wei Liu, Jianlong Zhou, Ren Ping Liu, Fang Chen
- Submitted: 15 September 2026
- arXiv identifier: arXiv:2609.17631 (cs.AI)
- Subjects: Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR); Computers and Society
- Length: Main article 24 pages, 4 figures; supplementary material 13 pages
- Status: Preprint; not peer-reviewed
- Corresponding author: Yifei Dong
Each obligation yields a target-bound witness, a localized counterexample, or a localized unverifiability โ and none can compensate for another. Only a fresh, complete all-pass record derives the permit consumed by one atomic publication transition.
Evaluation Methodology
The study employs identity vectors, adversarial cases, finite models, and historical implementations:
Results and Limitations
The results support internal coherence, bounded safety, fault sensitivity, and limited constructibility. However, they do not support claims of factual truth, general refinement, blind interoperability, field efficacy, or standards status.
Publication Details
via ArXiv AI