A senior Crown Prosecution Service prosecutor cited a fictional legal case in high court proceedings concerning the prosecution of a boy accused of raping his younger sister. The Guardian understands that the false reference appeared in an AI-generated response to an internet search, raising further concerns about unreliable material entering the justice system.
Court documents show that Ravinder Dhillon, a senior crown prosecutor in the CPS rape and serious sexual offences unit, submitted the citation while defending the decision to prosecute. The boy was 11 at the time of the alleged offence, and his sister was seven. His lawyers challenged the decision, arguing that prosecutors had failed to consider representations from social workers, youth workers and a forensic psychologist.
Dhillon decided in April that the prosecution should proceed, a position subsequently confirmed by the chief crown prosecutor. In October, she acknowledged that the purported 2024 case she had referenced did not exist. She said she had intended to refer to a genuine decision made 15 years earlier. Two senior high court judges described her explanation as unsatisfactory.
The high court said the false citation had not affected the outcome of the boy’s challenge to the continuation of his prosecution. His trial is scheduled for the youth courts later in the week of 3 November. Brian Flanagan, an associate professor of law at Maynooth University in Ireland who researches judicial use of AI, called reliance on an invented authority to support prosecuting a child a serious failure.
Flanagan said fabricated authorities undermined the trust placed in legal sources and that the time savings offered by AI increased the risk of such errors. He said the issue carried particular significance when the professional involved represented the state, given its considerable power.
This is the second known incident involving the CPS, the principal criminal prosecution agency in England and Wales. In February, it acknowledged that a lawyer in an extradition case concerning two Romanians convicted of minor offences had cited nonexistent authorities, probably because of AI. Lawyers on the opposing side discovered the problem when they could not locate the cases. The CPS subsequently assessed the likelihood of a repeat as low.
Responding to the latest incident, the CPS said its strict AI policy prevents the technology from making legal decisions and requires people to check accuracy. A spokesperson attributed the error to internet-based legal research and said it was corrected once discovered. An internal review examined how it happened and checked whether other cases had been affected. The agency said it then shared lessons, including the requirement to obtain case law from trusted legal sources.
Courts have encountered dozens of AI-related errors this year, largely involving people representing themselves without legal qualifications and frequently arising in employment tribunals. In June 2025, the high court urged the Bar Council and Law Society to address the problem urgently. Dame Victoria Sharp, then president of the King’s bench division, warned that misuse could damage justice and public confidence, with sanctions potentially including public admonishment, contempt proceedings or police referral.
The incident takes a UK tracker maintained by Doughty Street Chambers barrister Matthew Lee to 101 entries covering confirmed, suspected and disputed AI errors and related false citations. Lee warned that unpublished lower court decisions could conceal more examples. He also highlighted confidentiality risks involving sensitive CPS information and the additional work barristers face in verifying authorities in papers potentially prepared using AI.
