[
  {
    "type": "Preprint",
    "title": "Answer-Key Burden: A State-Responsibility Framework for Long-Horizon Human–AI Collaboration",
    "summary": "Proposes a state-responsibility framework for identifying human recovery work in long-horizon human–AI collaboration. The construct and proposed measurement approach are not established measures of human benefit.",
    "date": "2026-08-22",
    "status": "Zenodo v0.1 • DOI 10.5281/zenodo.22062448",
    "tags": [
      "human–AI collaboration",
      "state responsibility"
    ],
    "url": "https://doi.org/10.5281/zenodo.22062448",
    "external": true
  },
  {
    "type": "Preprint",
    "title": "When the Evaluation Pipeline Fails Before the Model: Failure, Repair, and Claim Ceilings in Multi-Stage AI Evaluation",
    "summary": "Examines failure ownership and repair when parts of an evaluation pipeline fail before the target is fully measured. Introduces a provisional method for preserving partial evidence without overstating conclusions.",
    "date": "2026-08-25",
    "status": "Zenodo v1 • DOI 10.5281/zenodo.22091682",
    "tags": [
      "AI evaluation",
      "failure ownership"
    ],
    "url": "https://doi.org/10.5281/zenodo.22091682",
    "external": true
  },
  {
    "type": "Preprint",
    "title": "Evaluator Provenance in Multi-Stage AI Evaluation: Separating Repeatability, Agreement, Human Ratification, and Decision Authority",
    "summary": "Separates repeatable evaluator outputs, agreement among evaluators, human ratification and the authority to make a decision. Publication of the preprint does not establish independent validation.",
    "date": "2026-08-31",
    "status": "Zenodo v1 • DOI 10.5281/zenodo.22207383",
    "tags": [
      "evaluator provenance",
      "decision authority"
    ],
    "url": "https://doi.org/10.5281/zenodo.22207383",
    "external": true
  },
  {
    "type": "Preprint",
    "title": "When the Judge Becomes the Experiment",
    "summary": "Studies evaluator behavior as a measurement variable, including the difference between reproducibility, agreement and validity. Bounded findings do not establish general model superiority.",
    "date": "2026-09-05",
    "status": "Zenodo v1.0 • DOI 10.5281/zenodo.22340649",
    "tags": [
      "evaluation validity",
      "model-as-judge"
    ],
    "url": "https://doi.org/10.5281/zenodo.22340649",
    "external": true
  },
  {
    "type": "Preprint",
    "title": "Scorer Defects After the Result",
    "summary": "Investigates scoring defects discovered after an evaluation result and their implications for reporting, recomputation and the conclusions supported by the original run.",
    "date": "2026-09-05",
    "status": "Zenodo v1.0 • DOI 10.5281/zenodo.22343213",
    "tags": [
      "scorer integrity",
      "evaluation repair"
    ],
    "url": "https://doi.org/10.5281/zenodo.22343213",
    "external": true
  },
  {
    "type": "Preprint",
    "title": "The Perfect Score With a Missing Case",
    "summary": "Examines how incomplete evaluation coverage can coexist with an apparently complete score, and why case accounting must be preserved before interpreting performance.",
    "date": "2026-09-06",
    "status": "Zenodo v1.0 • DOI 10.5281/zenodo.22541029",
    "tags": [
      "coverage integrity",
      "evaluation validity"
    ],
    "url": "https://doi.org/10.5281/zenodo.22541029",
    "external": true
  },
  {
    "type": "Preprint",
    "title": "When Translation Is Not Transition",
    "summary": "Examines the difference between translating information into another representation and demonstrating an intended state transition. Measured findings remain scoped to their evaluated cases.",
    "date": "2026-09-06",
    "status": "Zenodo v1.0 • DOI 10.5281/zenodo.22543624",
    "tags": [
      "state transition",
      "AI evaluation"
    ],
    "url": "https://doi.org/10.5281/zenodo.22543624",
    "external": true
  },
  {
    "type": "Preprint",
    "title": "When the Human Is the Hidden Answer Key",
    "summary": "Develops a protocol for evaluating recovery responsibility and answer-donation contamination in long-horizon AI work. This is a proposed method, not evidence of reduced human burden.",
    "date": "2026-09",
    "status": "Zenodo • DOI 10.5281/zenodo.22681331",
    "tags": [
      "human–AI evaluation",
      "answer donation"
    ],
    "url": "https://doi.org/10.5281/zenodo.22681331",
    "external": true
  }
]
