QUESTION PERIOD — Immigration, Refugees and Citizenship
Immigration Applications
October 30, 2024
Senator Gold, since 2014, Immigration, Refugees and Citizenship Canada, or IRCC, in order to boost efficiency, has been relying increasingly on artificial intelligence, or AI, algorithms in immigration processing. Senators on the Standing Senate Committee on Social Affairs, Science and Technology, or SOCI, have raised concerns about transparency, data security and innate discriminatory biases in AI use. Many Canadians share our concerns about skewed assessments discriminating on grounds such as country of origin, gender or socio-economic status. In the U.K., a similar AI visa process was discontinued when bias was proven.
Senator Gold, are legitimate applications to Canada, including study permits for Afghan women and girls endangered by gender apartheid, being denied by bots or by humans?
Thank you for your question and for raising an increasingly obvious point: Any system that uses a computer, in some sense, is also going to be subject to the algorithms that are either embedded in it or created for it.
I’m not aware of the specifics of your question, but the extent to which any of our systems, however managed — even those used by humans to evaluate files — contain biases, embedded or not, is a serious matter, and I’ll certainly bring this to the attention of the minister.
Thank you, Senator Gold.
I think we can agree that opaque decision making makes it difficult — sometimes impossible — to challenge discriminatory decisions.
Given Canada’s commitment to a fair immigration system, how is the Government of Canada ensuring oversight and transparency so that IRCC is not discriminatory?
The challenge of rooting out embedded or systemic structured biases that can result in discrimination in decision making is a perennial one. I won’t say it’s more difficult in an AI or algorithm world than it was in earlier times, but it certainly isn’t getting any easier to do that.
Again, I will bring this question to the attention of the minister. Thank you.
