UK prosecutor cited a fake AI-generated court case in a child rape trial
A senior Crown Prosecution Service lawyer relied on a non-existent legal precedent produced by an AI internet search — the second such CPS error this year.
What happened: Ravinder Dhillon, a senior Crown Prosecution Service lawyer, cited a legal precedent to the high court that turned out not to exist, while arguing that an 11-year-old boy should be prosecuted for allegedly raping his 7-year-old sister. The case, which she labelled 'CPS - R v J [2024] EWCA Crim 3529', came from an AI-generated answer during an internet search. Two high court judges called her explanation 'unsatisfactory'. The court said the fake citation did not change the outcome, and the boy's case is going to trial this week.
Why it matters: This is only the second known CPS case this year involving a fabricated citation likely caused by AI, after a February extradition case. An expert in AI and law called it a 'serious failure', noting that courts rely on precedents being verifiably real, and that the stakes rise sharply when the fabrication comes from the state's own prosecutor in a case involving a child. It adds to a UK-wide tracker of at least 101 confirmed, suspected or disputed AI-related errors in court filings this year, a count one barrister says is likely just the 'tip of the iceberg'.
How it works, plainly: Rather than using a dedicated legal research tool, the error reportedly came from an ordinary internet search that returned an AI-generated summary. These summaries can sound authoritative while inventing case names, citations and outcomes that have no basis in real law. Unless someone manually checks the underlying legal database, a convincing but entirely fictional precedent can slip into a court filing undetected, exactly as happened here.
The rollout: The CPS says it has a strict AI policy barring AI from making legal decisions and requiring human checks for accuracy, and that it reviewed the case once the error surfaced, concluding case law must come from trusted legal sources. But this is the second breach of that policy found this year. In mid-2025, senior judges urged the Bar Council and Law Society to act 'urgently', warning lawyers who misuse AI could face sanctions up to contempt of court or police referral. Enforcement so far has not stopped repeat errors.