{"resourceId":"california-appeal-gillespie-chatgpt-expert-report-diversion-20261105","versions":[{"version":"external-91b2d916fee5127d88059113c71f8f285a9b857cbb543249b39610b40cbe6806","resource":{"id":"california-appeal-gillespie-chatgpt-expert-report-diversion-20261105","title":"California Fourth District Court of Appeal finds no abuse of discretion admitting forensic psychologist testimony prepared with ChatGPT used as an organizational “thought partner”","organization":"California Court of Appeal, Fourth District, Division One","sector":"Courts / AI-assisted expert evidence in criminal diversion","geography":"San Diego County / California, United States","publishedAt":"California Courts Newsroom, Nov. 5, 2026 (opinion filed Nov. 4, 2026; unpublished)","publicationDate":"2026-11-05","eventDate":"2026-11-04","sourceName":"California Courts Newsroom","sourceLabel":"C.A.: No Error in Allowing Expert to Testify About Report Generated With Help of AI","sourceUrl":"https://newsroom.courts.ca.gov/news/ca-no-error-allowing-expert-testify-about-report-generated-help-ai","evidenceClass":"government-evaluation","outcomeClass":"emerging","topics":["governance-procurement","knowledge-work","operating-model"],"finding":"On Nov. 5, 2026 the California Courts Newsroom reported that Division One of the Fourth District Court of Appeal held a trial judge did not abuse her discretion by allowing a forensic psychologist’s testimony after she admitted using ChatGPT as a “thought partner” to prepare a report opposing pretrial mental health diversion, where the judge credited that the tool was used for organizational purposes. The underlying unpublished opinion in Gillespie v. Superior Court (D088222), filed Nov. 4, 2026, denied writ relief. The defense argued AI assistance required Kelly/Frye proof and undermined independence; the appellate court deferred to the trial court’s credibility finding that AI did not impair the independence of the expert opinion, while agreeing in principle that AI use by a forensic expert can raise novel-technique concerns. The expert testified she used ChatGPT under a setting that excluded her material from the training model, for organization/polish, and denied using it to form the nexus opinion. The opinion is unpublished and non-citable under California Rules of Court rule 8.1115(a).","sledRelevance":"U.S. trial and appellate courts are starting to police AI use in expert reports through credibility and disclosure, not categorical exclusion—relevant to prosecutors, defenders and court IT policies on generative AI in casework.","evidence":"Opened the California Courts Newsroom item via HTTPS curl and converted to text. Also downloaded and read the full unpublished PDF opinion D088222 from courts.ca.gov (pdftotext) for holdings on ChatGPT use, Kelly/Frye arguments and the trial court’s credibility finding.","architectureImplications":"Expert workflows that touch ChatGPT need enterprise/no-train settings and prompt retention if challenged.","governanceImplications":"Courts should require experts to disclose AI assistance early enough for credibility testing, even before formal rule changes.","securityPrivacyImplications":"Putting case records into a generative AI tool raises privacy objections even when the court admits the testimony.","caveats":"Opinion is unpublished and non-citable; fact-specific credibility finding; does not create a general California rule authorizing ChatGPT in expert work; diversion denial also rested on substantial-evidence grounds beyond AI.","streamIds":["public-safety"],"roles":{"sales":"Sales framing: California’s appellate newsroom story shows AI-in-expert-reports is a live courtroom risk, not a policy draft. Court and DA buyers need approved tools, disclosure scripts and training before the next challenge. Offer a bounded playbook for expert AI disclosure and no-train tooling, including how to document organizational versus analytical use. Do not claim Gillespie green-lights ChatGPT for all experts, that Kelly/Frye never applies, or that an unpublished opinion binds other districts.","engineering":"Engineering view: if experts use generative AI, prefer contracted no-train endpoints, logged prompts and human authorship gates. The Gillespie expert’s “thought partner” use still produced a factual error about medical records that defense highlighted on cross. Build checklists that force experts to re-read source records after any AI rewrite and to preserve prompts if a court later demands them. Do not treat consumer ChatGPT as court-safe by default.","delivery":"Delivery plan: issue interim guidance for retained experts, require AI-use declarations in reports, and pick an approved tool with a data-processing agreement. Dependencies are court leadership, DA and public-defender policies, and vendor contract terms. Acceptance is a written disclosure template used on the next three expert reports with a privacy-complaint path. Risks include privacy complaints, hidden consumer AI use, and overreading an unpublished opinion."},"roleDetails":{"sales":{"customerProblem":"Courts face challenges to AI-touched expert reports without clear rules.","stakeholders":["Judges","District attorney","Public defender","Court CIO"],"discoveryQuestions":["Do experts disclose AI today?","Is there an approved no-train tool?","Who owns the disclosure template?"],"valueHypothesis":"Early disclosure and approved tooling reduce exclusion fights.","potentialEngagement":"Expert-AI disclosure playbook for criminal courts.","unsupportedClaims":["That ChatGPT is generally admissible","That Kelly/Frye never applies","That unpublished opinions bind other courts"]},"engineering":{"fit":"Relevant to court/DA environments where experts draft reports with generative AI.","architectureIntegration":"Approved LLM endpoint, prompt logs, document vault that blocks consumer paste.","prerequisites":["Policy","Vendor DPA","Training"],"constraints":["Unpublished guidance","Privacy statutes","Expert independence"],"security":"No-train settings; minimize PHI/case facts in prompts; retain prompts if ordered.","proofOfValue":["Disclosure on file","No consumer-tool use","Accuracy check after AI edit"]},"delivery":{"work":"Draft disclosure rule; approve a tool; train retained experts; monitor challenges.","dependencies":["Court admin","Counsel","Vendor"],"ownership":"Presiding judge or court executive with DA/PD input.","skillsAdoption":"Experts must explain AI use under oath without over-delegating analysis.","governanceCheckpoints":["Policy published","First disclosed report","Post-hearing review"],"acceptanceCriteria":["Template used","Approved tool only","Privacy complaint path defined"],"risks":["Overreading unpublished case","Hidden consumer AI use","Privacy breach"]}},"retrievedAt":"2026-11-07T03:56:50Z","enrichedAt":"2026-11-07T04:05:20Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Defense and prosecution experts need shared disclosure norms so AI use does not become a surprise mid-hearing.","procurementImplications":"Court and DA offices may need approved AI tools with audit and no-train contracts rather than consumer ChatGPT.","operatingModelImplications":"Experts remain the authors; AI as organization/polish still faces cross-examination on independence and accuracy.","sourceVerification":{"openedUrl":"https://newsroom.courts.ca.gov/news/ca-no-error-allowing-expert-testify-about-report-generated-help-ai","referenceExcerpt":"a forensic psychologist who admitted that she used ChatGPT as a “thought partner” in preparing a report","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}