The Crown Prosecution Service (CPS) was forced to apologise after court documents prepared using artificial intelligence were found to contain references to 'non-existent' legal cases, a High Court judge has revealed.
Judge Highlights AI Risks in Legal Work
Mr Justice Sweeting made the comments on Wednesday while dismissing the appeals of two individuals challenging decisions to extradite them to Romania to serve sentences for offences committed there. The CPS opposed the appeals, but at the end of the ruling, the judge disclosed that two legal authorities cited by the CPS had been 'hallucinated' by an AI and did not actually exist.
The judge stated that the two fabricated cases were then cited in another document before the issue was raised with the CPS prior to a hearing in February. The CPS acknowledged and apologised for the error, stating that it had not sought to mislead the court and that inquiries were being undertaken into how the mistake occurred.
CPS Blames Human Error for Failing to Check AI Output
In a letter in March, the CPS accepted that the references to the cases were inaccurate and “explained that they were likely to have originated from the use of artificial intelligence,” the judge said. “However, the CPS identified the critical failing as the fact that the reviewing lawyer did not properly check the accuracy of the document before it was filed and served.”
He continued: “The position of the CPS was therefore that, while the immediate source of the error may have been the use of generative artificial intelligence, the operative cause was human error in the failure to verify the authorities relied upon in formal submissions placed before the Court. The CPS emphasised that this was not a deliberate attempt to mislead, but rather an isolated incident arising from inadequate checking of written work.”
Internal Review Finds No Other Cases Affected
Mr Justice Sweeting also said that the CPS had conducted a full internal review into the incident and reviewed 78 other cases linked to the same lawyer, which identified no concerns and led it to believe there was a low risk of it happening again. He concluded: “As far as the appeal was concerned, the errors came to light prior to the hearing and therefore had no impact on argument or the court’s judgment. I have accepted the apology given on behalf of the CPS and the assurance that there was no attempt to mislead.”
Judge Warns of Increasing AI Use Without Oversight
He continued: “It would be naive to assume that there will not be an increasing use of artificial intelligence in legal work in future; indeed, that may be both necessary and beneficial. The episode highlights the risks of its use without appropriate oversight, particularly for legal research.”
In a statement, a spokesperson for the CPS, which does not use AI in legal decisions or when determining charging outcomes, said: “We apologised to the court for this error made in legal submissions. As soon as this was identified, we carried out an internal review to establish how this occurred and to check no other cases were affected and that any learning was identified to prevent a recurrence. The judge has made clear the error had no impact on legal arguments or the court’s eventual judgment and accepted our apology.”
