QUESTION PERIOD — Immigration, Citizenship and Refugees
Artificial Intelligence
April 14, 2026
My question is for the Government Representative in the Senate.
A few weeks ago, in the midst of our debates over Bill C-12, our colleague Senator McPhedran asked a question at the National Security Committee about the use of AI in the processing of immigration requests. She was told that AI is not used to make those decisions.
In the last few weeks, however, we have seen disturbing stories about AI hallucinations poisoning people’s immigration applications. I wonder if you could clarify for us how AI is used in processing immigration applications and what guarantees we have that AI mistakes are not going to impede the applications of people who are otherwise qualified.
I reviewed that dossier quite recently, and it is my understanding that AI is not used to make any final decisions whatsoever concerning immigration. My understanding is that AI is used to sort the files themselves and help accelerate the process, but no decision making whatsoever concerning immigration is made by AI.
I guess the problem, though, is whether the sorting is corrupted by AI hallucinations — AI making up lies about people. I hate to use the word “hallucination” because it implies that AI is thinking. In fact, AI is putting words together like a word salad. So it’s garbage in, garbage out. If AI is triaging people, will it preclude them from having a human review their file to make sure there are no errors?
It’s my understanding that they are not triaging people. It’s used to help computerize the system. There is no decision whatsoever concerning immigration or asylum seekers that is deferred to AI. It always follows a due process with human intervention.
