The United Kingdom's asylum decision-making process has come under intense judicial scrutiny after a senior judge suggested that the Home Office may have relied on artificial intelligence-generated "hallucinated" information to reject the asylum application of a Moroccan woman and her child.

The case has sparked fresh concerns about the use of generative AI in high-stakes government decision-making, particularly where fundamental human rights and refugee protection are at stake.

Judge Finds Key Government Evidence Did Not Exist

The asylum claim was brought by a Moroccan woman who fled her home country with her child after allegedly enduring forced child marriage, rape, and years of severe domestic abuse. She argued that returning to Morocco would expose her to a real risk of being killed by her husband, whom she described as a powerful convicted criminal.

The Home Office rejected her application, relying on what it described as a Country Policy Information Note (CPIN) dated July 2021. According to the refusal letter, the document showed that Morocco offered sufficient protection for women in her circumstances, making her return safe.

However, during proceedings before the Upper Tribunal, the presiding judge discovered that no such CPIN could be located.

In his ruling, the judge stated that neither publicly available records nor the Home Office's own country policy and information team could confirm the existence of the cited document. He concluded that "it appears that no such CPIN has ever existed."

Possible AI 'Hallucination'

The judge went further by suggesting that the refusal letter displayed characteristics consistent with text generated by artificial intelligence.

He observed that the references to the nonexistent policy document "bear hallmarks consistent with the use of artificial intelligence," raising the possibility that the citation resulted from an AI "hallucination"—a phenomenon in which AI systems generate false or fabricated information that appears credible.

If this were the case, the judge described it as "an extremely serious failing" by the Home Office.

He also likened reliance on a document that never existed to relying on "bogus evidence," characterizing the issue as a serious procedural irregularity capable of undermining the integrity of the asylum decision.

Questions Over Home Office Evidence

When questioned about the missing document, the Home Office initially directed inquiries to an archive where the alleged CPIN could not be found.

The department later produced a different document, an unpublished internal Country Information Note (CIN), which had not been disclosed during the legal proceedings. Unlike the cited CPIN, the internal guidance did not conclude that Morocco was generally safe for women facing domestic violence or forced marriage. Instead, it acknowledged continuing concerns, including widespread child marriages and the absence of explicit criminalization of marital rape under Moroccan law.

The Upper Tribunal is continuing to examine the woman's appeal.

Home Office Responds

Responding to the controversy, a Home Office spokesperson did not directly address the judge's observations regarding artificial intelligence.

Instead, the department stated that the government is focused on restoring order to the UK's asylum system by increasing removals of individuals without legal permission to remain in the country while ensuring that every asylum application is assessed individually using the latest available evidence.

The woman's appeal remains before the courts, but the case has already intensified debate over the role of artificial intelligence in public administration. If the judge's concerns are ultimately substantiated, the ruling could have significant implications for the Home Office's use of AI in asylum determinations and may prompt wider scrutiny of automated decision-making across government agencies.

Source: The Guardian