Canada Immigration AI Hallucinated Job Duties — PhD Immunologist Denied Permanent Residency (2026)

Incident Summary

PhD immunologist Kemy Ade was denied permanent residency in Canada after Immigration, Refugees and Citizenship Canada’s (IRCC) AI system fabricated job duties in the refusal letter, describing her as “wiring control circuits, building robot panels” — duties that bear no relationship to immunology research.^[1] The case represents the first documented acknowledgment by IRCC that generative AI was used in immigration decision-making, revealing that Canada’s immigration agency had been deploying AI to process applications without public disclosure.^[2] The AI adoption was driven by a backlog of over 1 million applications, creating pressure to automate decisions at the cost of accuracy.^[3] The hallucinated job duties — entirely fabricated by the AI system and bearing no connection to the applicant’s actual qualifications — demonstrate that generative AI hallucinations in government decision-making can have direct, life-altering consequences for individuals, transforming a technical limitation of language models into a denial of legal status.

Key Facts

• Applicant: Kemy Ade, PhD immunologist^[1]
• AI fabrication: Described her as “wiring control circuits, building robot panels”^[1]
• Decision: Permanent residency denied based on hallucinated duties^[1]
• Precedent: First IRCC acknowledgment of generative AI in immigration decisions^[2]
• Context: 1 million+ application backlog driving AI adoption^[3]

Threat Patterns Involved

Primary: Misinformation & Hallucinated Content — The AI system generated completely fabricated job duties that were incorporated into an official government decision, demonstrating how LLM hallucinations in automated government systems produce authoritative-seeming misinformation with direct legal consequences.

Secondary: Overreliance & Automation Bias — The incorporation of AI-hallucinated job duties into an official refusal letter without human verification indicates that IRCC decision-makers treated the AI’s output as accurate without cross-referencing against the applicant’s actual documentation.

Significance

1. Government hallucination with legal force — The fabricated job duties were incorporated into an official immigration decision with legal standing, demonstrating that AI hallucinations in government systems carry consequences that extend beyond misinformation to the denial of legal rights
2. First acknowledged government AI in immigration — IRCC’s acknowledgment that generative AI was used in immigration decisions reveals previously undisclosed automation of consequential government decision-making
3. Backlog-driven automation risk — The 1 million+ backlog creating pressure for AI adoption demonstrates how administrative pressure can override the caution necessary for deploying AI in high-stakes decisions
4. PhD immunologist as test case — The absurdity of describing a PhD immunologist as “wiring control circuits” makes the hallucination obvious, but raises the question of how many less obviously wrong hallucinations have gone undetected in other immigration decisions