Thank you for the question.
It would really be in two parts. I think Canada generally needs AI legislation, and then we need legislation specific to immigration. The economic action plan that was passed in 2015 was seven years ago. It's on that very broad legislation, with very little detail, that all of this AI is being actioned, but there was not even a debate on it at the time. Although it was emerging, there were no comments.
When we speak about the companies and who's driving the technology, what's important to understand is third party use of technology like AI. IRCC, to their credit, are developing AI in-house, but other places, like the Canada Border Services Agency and Employment and Social Development Canada, are outsourcing to a third party. It's very easy at that point to begin to lose traction. Who has the business assets? Who has the proprietary control?
On general legislation, we can look at countries that are further ahead of us, such as New Zealand, England and Australia, and the measures they've put in place. In terms of legislation, I think we need to start right away with an algorithmic charter that sets out the types of algorithms. I noticed that during the debate there was a lot of discussion today about models, but the meat of everything is in the algorithms and, in Immigration Canada's case, the undisclosed officer rules. Those rules are not even disclosed to the officers who are deciding the applications, but that's the basis on which they're triaging, so we have a lot of work to do to get up to the ethical standards we've seen and the recommended governance standards we've seen throughout the world.