CPS Prosecutor Cited a Nonexistent Case to the High Court After an AI-Generated Search

Ravinder Dhillon, a senior crown prosecutor in the CPS rape and serious sexual offences unit, was defending the decision to prosecute a boy who was 11 at the time he allegedly raped his seven-year-old sister. The boy's lawyers had challenged that decision, arguing prosecutors failed to consider representations from social workers, youth workers and a forensic psychologist. In arguing for the prosecution, Dhillon submitted a detailed reference to a case she named as CPS – R v J [2024] EWCA Crim 3529. No such case exists.

Dhillon took the decision to proceed in April, and it was confirmed by the chief crown prosecutor. In October she admitted the citation had been made in error and said she had meant to refer to a different, real decision from 15 years earlier. Two senior high court judges described her explanation as "unsatisfactory", The Guardian reported. The court said the false reference did not change the outcome of the boy's challenge, and the case is listed for trial in the youth courts later this week.

The CPS Response

The CPS said it has a strict AI policy "which limits its use so that it is not used to make legal decisions and there must be human checks for accuracy". A spokesperson said the agency "identified an error had occurred from conducting legal research using the internet", carried out an internal review to check that no other cases were affected, and shared lessons internally, including that case law must be cited from trusted legal sources.

This is the second known case involving the CPS. In February the agency admitted that one of its lawyers had cited nonexistent authorities, likely produced by AI, in an extradition case involving two Romanians convicted of minor offences. Those citations were spotted only when opposing lawyers could find no trace of them, and the CPS concluded at the time that "the risk of recurrence was low".

A Growing Record of AI Errors in Court

Brian Flanagan, an associate professor of law at Maynooth University in Ireland who studies judges' use of AI, called the episode "a serious failure". He said the stakes are higher when the professional involved is acting for the state, "because the power of the state is immense."

The incident is the 101st entry in a UK tracker of confirmed, suspected and disputed AI-related errors and false citations kept by Matthew Lee, a barrister at Doughty Street Chambers. Lee said the recorded cases may be the "tip of the iceberg" because many lower court rulings are never published. He also pointed to privacy and confidentiality risks when an agency handling highly sensitive material uses AI, and said barristers now spend extra time checking whether authorities in case papers are genuine. Most of the errors logged this year came from people representing themselves, often in employment tribunals.

In June 2025 the high court asked the Bar Council and the Law Society to consider steps to curb AI hallucinations "as a matter of urgency". Dame Victoria Sharp, then president of the King's Bench Division, warned that lawyers who misuse AI could face sanctions ranging from public admonishment to contempt proceedings and referral to the police.

For legal departments and professional services firms, the case shows how AI can slip into research through everyday tools such as search engines, outside any formal deployment, and why verifying citations against trusted databases remains a basic control.