A UK tribunal judge has condemned the Home Office for potentially utilizing AI-generated,fictitious information to reject an asylum application from a woman and her children fleeing Morocco. The ruling suggests that the government relied on a non-existent policy document to justify the deportation of a family escaping extreme violence and forced underage marriage.
The non-existent CPIN that denied asylum
The legal dispute centers on the UK Home Office's reliance on a "country policy information note" (CPIN) that the tribunal judge ruled does not actually exist. According to the report,this phantom document was used as the primary justification for claiming that Morocco was a safe destination for the woman and her children to return to, despite the woman's traumatic history of violence.
The Upper Tribunal's hearing revealed that the UK Home Office was unable to produce or even confirm the existence of the CPIN in question, even after repeated demands from the legal teams. The judge characterized this as "relying on bogus evidence," suggesting that the reference to the document may have been the result of an AI-generated error, commonly known as a hallucination.
How the internal CIN revealed 20% child marriage rates
While the Home Office could not find the CPIN, it eventually produced an unpublished internal guidance document known as a "country information note" (CIN).. As reported, this internal CIN directly contradicted the government's previous assertions regarding the safety of Morocco, providing a starkly different picture of the risks facing women and children.
The internal CIN disclosed that child marriages accounted for 20% of all marriages in Morocco in 2018. Furthermore, the document highlighted a critical legal gap: Morocco lacks explicit laws against marital rape, and the report notes that women who attempt to report such crimes are often the ones who end up criminalized.
AI hallucinations in the UK Home Office's decision-making
The potenital use of AI-generated fabrications in a legal proceeding points to a systemic risk within the UK Home Office's administrative processes. This case echoes a growing global concern where government agencies integrate large language models into drafting workflows without sufficient human oversight, leading to "hallucinations" where the AI invents citations or legal precedents to fill gaps in its logic .
The judge expressed grave concerns regarding the misuse of AI in generating critical policy documents that determine the fate of vulnerable individuals. This incident suggests that the UK's asylum process may be suffering from a lack of rigorous verification, where algorithmic output is being accepted as factual evidence without being cross-referenced against actual policy archives.
Who authorized the use of unverified AI tools?
Significant questions remain regarding the Home Office's internal protocols, specifically whether the department has officially sanctioned the use of generative AI for drafting asylum decisions.. While the Home Office maintains that claims are assessed individually based on the latest evidence , it has not explained how a non-existent CPIN entered a formal legal decision.
It remains unclear how many other asylum seekers have had their claims denied based on similarly fabricated AI evidence. Because the Home Office has not disclosed the specific software or prompts used in this case, the full scale of the procedural irregularity remains hidden from the public and the legal community.
Comments 0