{"resourceId":"uk-cps-prosecutor-cites-nonexistent-ai-generated-case-in-high-court-child-prosecution-20261103","versions":[{"version":"external-66d819ce5d5508754ad0a9179f9f9d054983947337de2577c80474a530836da6","resource":{"id":"uk-cps-prosecutor-cites-nonexistent-ai-generated-case-in-high-court-child-prosecution-20261103","title":"A Crown Prosecution Service lawyer cited a nonexistent case, traced to an AI search summary, while arguing in the High Court to prosecute a child","organization":"Crown Prosecution Service (England and Wales)","sector":"Courts / prosecution / legal research with AI","geography":"England and Wales, United Kingdom","publishedAt":"The Guardian, Nov. 3, 2026","publicationDate":"2026-11-03","eventDate":"2026-10-31","sourceName":"The Guardian","sourceLabel":"A serious failure: CPS prosecutor cites imaginary case to high court after AI search","sourceUrl":"https://www.theguardian.com/technology/2026/nov/03/a-serious-failure-cps-prosecutor-cites-imaginary-case-to-high-court-after-ai-search","evidenceClass":"independent-reporting","outcomeClass":"cautionary","topics":["knowledge-work","governance-procurement","data-security"],"finding":"The Guardian reported that a senior crown prosecutor in the CPS rape and serious sexual offences unit cited a nonexistent precedent, \"R v J [2024] EWCA Crim 3529\", in a High Court challenge to the decision to prosecute a boy who was 11 at the time of an alleged offence. The prosecutor admitted the error in October; two High Court judges called her explanation unsatisfactory. The Guardian understands the citation came from an AI summary in an internet search. The court said the error did not change the outcome. It is the second known CPS case after a February extradition matter. The CPS says its policy bars AI from legal decisions and requires human accuracy checks; after an internal review it told staff case law must come from trusted legal sources. A barrister’s UK tracker now lists 101 AI-related court errors.","sledRelevance":"Shows a government prosecutor, not a self-represented litigant, introducing an AI-fabricated authority despite an existing AI policy; the control gap was in everyday research tools, not approved AI systems.","evidence":"Opened and read the full Guardian article. Court documents are described but not linked; the CPS internal review was not published.","architectureImplications":"The failure came from an AI summary inside a general web search, outside any approved tool, so tool allow-lists alone do not cover it.","governanceImplications":"An AI policy existed and still failed; citation verification against authoritative databases needs to be a checked step, not a policy statement.","securityPrivacyImplications":"A barrister warned of confidentiality risks when prosecutors handle sensitive case material near AI tools.","caveats":"UK prosecution and court practice differ from U.S. state systems; facts come from one news report and court documents it describes; the AI source is the newspaper’s understanding, while the CPS said internet research.","streamIds":["public-safety"],"roles":{"sales":"Sales framing: prosecutors and public defenders face the same risk the CPS just exposed: an AI summary inside an ordinary web search can plant a fake authority in a filing. The buyer problem is proving every citation is real before it reaches a judge. A bounded engagement could map where staff research, then add citation verification to filing workflows. Do not claim AI caused the error with certainty; the CPS described it as internet research.","engineering":"Engineering framing: the weak point was not an approved AI system but a search engine that returns generated summaries. Pre-sales should ask which research tools staff actually use, whether filings pass an automated citation check against an authoritative database, and how drafts are logged. A proof of value would run a citation validator across a sample of past filings and report unmatched authorities. Constraints include case-management integration and confidential material.","delivery":"Delivery framing: the work is adding a verification checkpoint to legal drafting in a prosecution or defense office. Ownership sits with the chief prosecutor or office lead, supported by legal research staff and IT. Acceptance criteria are proposals: every cited authority checked against a trusted source before filing, checks recorded, and staff trained on AI summaries in search. Main risks are time pressure, policy without enforcement, and errors surfacing only when opposing counsel looks."},"roleDetails":{"sales":{"customerProblem":"Government lawyers can file AI-fabricated authorities even under an AI policy, damaging cases and public trust.","stakeholders":["Chief prosecutor or district attorney","Public defender leadership","Legal research librarian","Court liaison"],"discoveryQuestions":["Which research tools do staff use day to day?","Is each citation checked against an authoritative database before filing?","How are errors reported and reviewed?"],"valueHypothesis":"A verified-citation workflow reduces sanctions risk and protects case outcomes.","potentialEngagement":"Citation-verification workflow and staff training for a prosecution or defense office.","unsupportedClaims":["That AI definitely produced the citation","That the error affected the case outcome","That a policy alone prevents recurrence"]},"engineering":{"fit":"Relevant to any prosecution, defense or court office drafting legal filings.","architectureIntegration":"Citation checking integrated with document drafting and case management, flagging authorities not found in a trusted database.","prerequisites":["Access to an authoritative case-law database","Inventory of research tools in use","Filing workflow with review step"],"constraints":["Confidential case material","Time pressure","Staff use of personal search tools"],"security":"Keep confidential case material out of public AI tools and log use of approved ones.","proofOfValue":["Run a validator on past filings","Measure unmatched citations","Time added per filing"]},"delivery":{"work":"Add a citation-verification step and training to the drafting process.","dependencies":["Legal research subscription","Case-management integration","Supervisor sign-off"],"ownership":"Office lead with legal research staff and IT.","skillsAdoption":"Staff need to recognise AI summaries in search results and verify every authority.","governanceCheckpoints":["Verification rule adopted","Pilot on one unit","Review of errors after 90 days"],"acceptanceCriteria":["All filings carry verified citations","Checks recorded","No fabricated authority filed during pilot"],"risks":["Policy without enforcement","Workload pressure","Errors found only by opponents"]}},"retrievedAt":"2026-11-04T03:49:23Z","enrichedAt":"2026-11-04T03:52:08Z","enrichmentBasis":"retrieved source","accessibilityWorkforceImplications":"Opposing counsel and barristers now spend time verifying every authority, shifting workload across the system.","procurementImplications":"Legal research subscriptions with citation checking may be a cheaper control than banning AI; agencies should budget for verification tooling.","operatingModelImplications":"Prosecution offices need a review step where a second person or a tool confirms each cited authority before filing.","sourceVerification":{"openedUrl":"https://www.theguardian.com/technology/2026/nov/03/a-serious-failure-cps-prosecutor-cites-imaginary-case-to-high-court-after-ai-search","referenceExcerpt":"Following the rapid review we shared learnings, including that case law must be cited from trusted legal sources.","promptVersion":"sled-research-v4.0","model":null,"basis":"agent-reported inspection"}}}]}