{
  "researchId": "JPC-26-00136",
  "researchUrl": "https://jipcet.org/r/JPC-26-00136",
  "articleUrl": "https://jipcet.org/articles/jipcet-2026-0136",
  "articleRef": "JIPCET-2026-0136",
  "externalDoi": null,
  "title": "Trust Calibration in Human–AI Decision Support: Confidence Displays, Failure-Mode Explanation and Appropriate Reliance",
  "authors": [
    {
      "name": "Olivia Whitmore",
      "affiliation": "University of Essex, United Kingdom",
      "orcid": null
    },
    {
      "name": "Hannah Beaumont",
      "affiliation": "Toronto Metropolitan University, Canada",
      "orcid": null
    },
    {
      "name": "Marcus Doyle",
      "affiliation": "University of Essex, United Kingdom",
      "orcid": null
    }
  ],
  "abstract": "Background. Decision-support interfaces routinely display model confidence on the assumption that numerical uncertainty enables users to rely on advice selectively. Evidence that confidence alone improves the quality of reliance decisions is mixed. Objective. We test whether confidence displays support appropriate reliance — accepting correct advice and rejecting incorrect advice — and whether pairing confidence with an explanation of the model's characteristic failure modes changes that relationship. Methods. Two pre-registered within-subjects experiments (total n = 214; Study 1 n = 96 clinicians and nurses, Study 2 n = 118 insurance adjudicators) used a triage task with 48 cases per participant. We manipulated confidence display (absent, numeric, verbal) and failure-mode explanation (absent, present), holding model accuracy at 0.78 with a stratified error distribution. Primary outcomes were appropriate-reliance rate and a signal-detection measure of reliance discrimination; we also recorded decision time and self-reported trust. Results. Numeric confidence alone raised overall agreement with the model from 0.61 to 0.74 but did not improve reliance discrimination, because agreement rose on incorrect advice as well as correct advice — over-reliance rather than calibration. Adding failure-mode explanation improved discrimination substantially and reduced agreement on incorrect high-confidence advice by 19 percentage points. The interaction between confidence and explanation was significant in both studies and larger among more experienced participants. Explanation added a mean 4.1 seconds per decision. Conclusion. Confidence information is not self-interpreting. Interfaces should pair uncertainty with an account of when the model is characteristically wrong; displaying confidence without it measurably increases misplaced reliance.",
  "keywords": [
    "human-AI interaction",
    "trust calibration",
    "decision support",
    "appropriate reliance",
    "explainability",
    "uncertainty communication"
  ],
  "area": "Human–Technology",
  "issue": "JIPCET 2026; 3(1)",
  "pageRange": "5–34",
  "published": "2026-03-12",
  "received": "Received 02 Oct 2025",
  "version": "1.0",
  "versions": [
    {
      "version": "1.0",
      "date": "2026-03-12",
      "note": "Version of record published after editorial decision."
    }
  ],
  "articleType": "Research Article",
  "reviewStatus": "Peer reviewed under named human editorial supervision",
  "peerReviewed": true,
  "handlingRecord": "Handled by a named JIPCET editor of record, who read the manuscript, oversaw review and approved the decision.",
  "methodology": "Research Article",
  "dataAvailability": "Available from the corresponding author on reasonable request, in line with participant and organisational agreements.",
  "codeAvailability": "Analysis procedures are documented in full in the methods section of the article.",
  "referenceCount": 6,
  "references": [
    "Whitmore, O. (2025). Reliance and calibration in clinical decision aids. Human Factors 67(2), 210–229.",
    "Beaumont, H. & Doyle, M. (2025). Discrimination measures for reliance research. ACM Transactions on Computer–Human Interaction 32(4), 1–29.",
    "Doyle, M. (2024). Trust attitudes and reliance behaviour: a dissociation. Ergonomics 67(9), 1188–1205.",
    "Beaumont, H. & Whitmore, O. (2025). Surface completeness and reviewer judgement. Human–Computer Interaction 40(4), 388–412.",
    "Paxton, D. & Okonkwo, R. (2026). Failure taxonomies as interface content. JIPCET 3(2), 1–40.",
    "Hale, E. (2025). Uncertainty communication in operational settings. Safety Science 181, 106–124."
  ],
  "replications": "No independent replication recorded to date.",
  "corrections": "No corrections issued. Any correction is published as a new version with a dated entry in the version history.",
  "license": "CC BY 4.0",
  "fullTextHtml": "https://jipcet.org/articles/jipcet-2026-0136",
  "fullTextPdf": null,
  "publisher": {
    "name": "JIPCET Editorial Office",
    "url": "https://jipcet.org",
    "email": "editor@jipcet.org"
  },
  "journal": {
    "title": "Journal of Innovation, Product, Computing & Emerging Technologies",
    "short": "JIPCET"
  }
}