A senior crown prosecutor representing the Crown Prosecution Service referenced a fabricated legal precedent while arguing for the prosecution of an 11-year-old child accused of rape. Ravinder Dhillon, acting in the CPS’s rape and serious sexual offences unit, submitted detailed citation to a case identified as CPS – R v J [2024] EWCA Crim 3529 during youth court proceedings. The citation was later revealed to be entirely fictitious.
Dhillon acknowledged the error in October, explaining that she had intended to reference a different, legitimate decision from approximately 15 years prior. Two senior high court judges characterized the CPS’s reliance on the nonexistent case as “unsatisfactory,” while Brian Flanagan, an associate professor of law at Maynooth University who studies AI use in legal contexts, described it as “a serious failure.” Flanagan emphasized that fabricating legal authorities violates fundamental professional norms, noting that the involvement of a state agency amplifies concerns about the misuse of governmental power.
The incident represents only the second confirmed case involving the CPS acknowledging AI-related citation errors. The agency previously admitted in February that one of its lawyers cited nonexistent legal authorities in an extradition matter, a situation the CPS assessed as presenting low recurrence risk. Investigation revealed that the false citation originated from an AI-generated internet search summary containing fictional case details. The CPS subsequently conducted an internal review and implemented corrective measures, including a requirement to cite case law exclusively from verified legal sources.
The high court determined that the fictional reference did not alter the outcome of the child’s challenge to the prosecution decision. Earlier in June 2025, senior judiciary had warned of serious implications for public confidence in the justice system resulting from AI misuse in legal proceedings. A tracker maintained by a barrister at Doughty Street Chambers documents 101 confirmed, suspected, and disputed AI-related legal errors, with observers suggesting this figure represents only a fraction of actual incidents given that many lower court decisions remain unpublished.
Article Attribution | Read More at Article Source
Article summary produced by Claude AI