{
  "version": 1,
  "updatedAt": "2026-09-05",
  "canonicalUrl": "https://cognitive-biases.github.io/research/changes/",
  "semantics": {
    "strengthens": "A useful new result supports an existing evidence record without making it universal.",
    "narrows": "A boundary condition, null result, or competing finding makes an existing claim more precise.",
    "new context": "Evidence moves a known decision problem into a practical setting that deserves explicit guidance."
  },
  "changes": [
    {
      "id": "2026-08-ai-medical-bias-review",
      "digest": "2026-08",
      "month": "August 2026",
      "publishedAt": "2026-09-05",
      "digestUrl": "https://cognitive-biases.github.io/research/digests/2026-08/",
      "signalId": "ai-medical-bias-review",
      "title": "AI-assisted medical decisions show many bias risks — and a domain-evidence gap",
      "changeType": "new context",
      "confidence": "moderate",
      "finding": "A structured review synthesized 24 studies and identified 12 cognitive biases reported across AI-assisted medical decision making. Only one included study came directly from pathology, the review's target domain.",
      "whyItMatters": "A long list of possible biases can look more settled than the underlying evidence. This review is useful partly because it exposes that gap: evidence from related medical settings can guide questions, but it should not be presented as direct pathology evidence.",
      "practicalCheck": "Ask two questions separately: has this bias been observed in human-AI decisions, and has it been observed in this exact task and professional setting?",
      "projectChange": "Add a domain-transfer caution to the AI-assisted decisions guide: related-domain evidence can motivate a check without proving the same effect in the target workflow.",
      "sourceStatus": "peer-reviewed structured review",
      "source": {
        "title": "Cognitive biases in AI-assisted medical decision making: A structured review as a primer for veterinary and human pathology",
        "publishedAt": "2026-08-05",
        "doi": "10.1177/03009858261472493",
        "url": "https://pubmed.ncbi.nlm.nih.gov/42557856/"
      }
    },
    {
      "id": "2026-08-illusory-truth-opinion-boundary",
      "digest": "2026-08",
      "month": "August 2026",
      "publishedAt": "2026-09-05",
      "digestUrl": "https://cognitive-biases.github.io/research/digests/2026-08/",
      "signalId": "illusory-truth-opinion-boundary",
      "title": "Repeating an opinion is not the same test as repeating a factual claim",
      "changeType": "narrows",
      "confidence": "moderate",
      "finding": "Two preregistered experiments with 457 participants found no reliable increase in truth ratings when social-political opinion statements were repeated. Equivalence tests indicated that any repetition effect in the tested conditions was smaller than the researchers' predefined meaningful threshold.",
      "whyItMatters": "The Illusory Truth Effect has a robust overall evidence base, but the popular version is often too broad. A checkable factual statement and an evaluative social-political opinion should not be assumed to respond to repetition in the same way.",
      "practicalCheck": "Before applying the Illusory Truth label, classify the statement. For factual claims, trace independent evidence. For evaluative opinions, also consider persuasion, identity, norms and preference rather than forcing a factual truth-judgment explanation.",
      "projectChange": "Narrow the canonical Illusory Truth page, publish a boundary-condition research note, and keep both the positive social-media result and the opinion null result visible in the same digest.",
      "sourceStatus": "peer-reviewed; two preregistered experiments and mini meta-analysis",
      "source": {
        "title": "Limits of the illusory truth effect for social-political opinions: Evidence from two experiments and a mini meta-analysis",
        "publishedAt": "2026-08-27",
        "doi": "10.1016/j.concog.2026.104119",
        "url": "https://doi.org/10.1016/j.concog.2026.104119"
      }
    },
    {
      "id": "2026-08-illusory-truth-social-media",
      "digest": "2026-08",
      "month": "August 2026",
      "publishedAt": "2026-09-05",
      "digestUrl": "https://cognitive-biases.github.io/research/digests/2026-08/",
      "signalId": "illusory-truth-social-media",
      "title": "Repetition stayed powerful in an Instagram-like setting",
      "changeType": "strengthens",
      "confidence": "moderate",
      "finding": "In a study of 165 participants, repeated statements were more likely to be judged true and were judged with greater confidence. Showing like counts did not meaningfully remove the repetition effects. Like magnitude mattered more for new statements, where repetition was not available as a cue.",
      "whyItMatters": "The classic repetition effect is not limited to plain laboratory lists. In this experiment, a social-media-like interface added another cue, but repetition still strongly shaped truth and certainty judgments.",
      "practicalCheck": "Before treating familiarity as evidence, count independent sources. Ten exposures to the same claim can still be one piece of evidence repeated ten times.",
      "projectChange": "Strengthen the social-media boundary condition on the Illusory Truth Effect page and add a practice item that separates source independence from exposure count.",
      "sourceStatus": "peer-reviewed experiment",
      "source": {
        "title": "The robustness of repetition-based illusory truth and certainty effects in social media contexts",
        "publishedAt": "2026-08-05",
        "doi": "10.1038/s41598-026-61449-y",
        "url": "https://pmc.ncbi.nlm.nih.gov/articles/PMC13443530/"
      }
    },
    {
      "id": "2026-08-llm-human-assessment-boundary",
      "digest": "2026-08",
      "month": "August 2026",
      "publishedAt": "2026-09-05",
      "digestUrl": "https://cognitive-biases.github.io/research/digests/2026-08/",
      "signalId": "llm-human-assessment-boundary",
      "title": "Not every bias manipulation works in every human-AI task",
      "changeType": "narrows",
      "confidence": "provisional",
      "finding": "A mixed-methods study compared 35 professionals with seven LLMs in neurodevelopmental support-allocation judgments. Neither group showed significant susceptibility to the study's anchoring or representativeness manipulations, while both showed inconsistencies between descriptive ratings and final decisions. LLMs reported higher intellectual humility, but that self-report was not related to decision consistency.",
      "whyItMatters": "This is a useful negative result for a bias library. A familiar label should not be forced onto every decision problem. The task may reveal a different failure mode than the one the experiment expected.",
      "practicalCheck": "Measure the decision change caused by the manipulation. Do not infer a bias from a plausible story about why the answer looks wrong.",
      "projectChange": "Use this as a boundary-condition example in the AI research methodology: bias labels require task-level evidence, and null results are informative.",
      "sourceStatus": "preprint",
      "source": {
        "title": "Language Models Reproduce Human Reductionist Bias and Decision Inconsistency in Neurodevelopmental Disorders Assessment",
        "publishedAt": "2026-08-17",
        "url": "https://arxiv.org/abs/2608.17105"
      }
    },
    {
      "id": "2026-08-llm-judge-prior-score-anchor",
      "digest": "2026-08",
      "month": "August 2026",
      "publishedAt": "2026-09-05",
      "digestUrl": "https://cognitive-biases.github.io/research/digests/2026-08/",
      "signalId": "llm-judge-prior-score-anchor",
      "title": "A previous score can pull an LLM judge toward it",
      "changeType": "new context",
      "confidence": "provisional",
      "finding": "Across 192,000 attempted evaluations of fixed texts, seven of eight tested models showed a task-bootstrap interval below zero for the total anchored-metadata effect. In a separate categorical dataset, anchored metadata blocked 48% of error corrections and changed 10.18% of correct judgments toward an assigned wrong label.",
      "whyItMatters": "Many agent loops keep previous scores, revision numbers, or reviewer comments in context. If the goal is an independent second judgment, that history can become part of the decision instead of harmless metadata.",
      "practicalCheck": "When you want an independent re-evaluation, hide the previous score first. Compare a blind judgment with a history-aware judgment instead of assuming they are equivalent.",
      "projectChange": "Add an LLM-as-a-judge scenario to AI-assisted reasoning practice and treat prior-evaluation metadata as a testable anchoring pathway, not as a universal model trait.",
      "sourceStatus": "arXiv manuscript; paper states CIKM 2026 publication",
      "source": {
        "title": "Anchoring Bias in LLM-as-a-Judge Systems: Prior Scores Compromise Evaluation Independence",
        "publishedAt": "2026-08-26",
        "url": "https://arxiv.org/abs/2608.25869"
      }
    }
  ]
}