Evidence of meeting #19 for Access to Information, Privacy and Ethics in the 44th Parliament, 1st Session. (The original version is on Parliament’s site, as are the minutes.) The winning word was frt.
A recording is available from Parliament.
4 p.m.
Liberal
4 p.m.
Conservative
The Chair Conservative Pat Kelly
Indeed. Thank you.
Mr. Villemure, you now have the floor for six minutes.
4 p.m.
Bloc
René Villemure Bloc Trois-Rivières, QC
Thank you, Mr. Chair.
My questions will be for Mr. Larter.
Mr. Larter, you said that Microsoft does not sell its technology to law enforcement. Does Microsoft have military agencies, surveillance agencies, or intelligence agencies as clients?
4 p.m.
Director, Responsible Artificial Intelligence Public Policy, Microsoft
We do think that facial recognition can have applications in the security and law enforcement context. I think what's really important to note here is that we take a risk-based approach to how we assess using a system and the kinds of customers that we will work with.
We have our sensitive use review process that I mentioned, with those three triggers. Anytime we're going to deploy a system—whether it's facial recognition or something else—in a way that will hit one of those three triggers, we go through a robust sensitive use review process.
We do make our facial recognition available for law enforcement and security uses. We think technology can have some useful applications in those scenarios, but it's really important that there be robust safeguards around that use, including the internal safeguards I've mentioned at Microsoft, but also a regulatory framework that ensures that we're all clear on how the technology is being used and know that it's being used in a way that is trustworthy and responsible.
4 p.m.
Bloc
René Villemure Bloc Trois-Rivières, QC
Could you, at a later date and in writing, provide the committee with examples of the safeguards in question?
4 p.m.
Director, Responsible Artificial Intelligence Public Policy, Microsoft
Yes, for sure. I would be very happy to do that.
4 p.m.
Bloc
René Villemure Bloc Trois-Rivières, QC
That's fine. Thank you very much.
The question may seem strange to you, but is Microsoft looking for this to be regulated so that later they can say they even advocated for a framework, and it's okay if certain things happen?
It may seem irrelevant, but does having such a framework protect Microsoft?
4 p.m.
Director, Responsible Artificial Intelligence Public Policy, Microsoft
It protects Microsoft to have a framework in place, but not necessarily for the reasons that were mentioned. We generally think that it's really important to build a regulatory framework for technology in general that engenders trust and shows that technology is being used in a trustworthy fashion.
We have been around for a while now. We're almost 50 years old as a company, and we realize that if society is going to reap the benefits of technology and if people are going to use it, they need to trust it. Regulation is a really important part of building that trustworthiness framework. That's what we advocate for in general, and that's particularly why we are investing time in trying to advocate robust safeguards around facial recognition, given that it is a very powerful technology with some very positive applications, as I mentioned, but potentially some challenges as well.
Creating a framework around facial recognition that ensures it can be used in a trustworthy way, and in a way the public sees is trustworthy as well, is very important for society so that it can reap the benefits and make sure that this technology is used over the longer term.
May 5th, 2022 / 4 p.m.
Bloc
René Villemure Bloc Trois-Rivières, QC
I really like your approach, by the way.
We would be grateful if you could send us, afterwards, any information you have on safeguards, types of programs and facial recognition technology liability, or anything related to that.
My next question may seem somewhat surprising to you.
Is there a connection between facial recognition technology and the new metaverse?
4 p.m.
Director, Responsible Artificial Intelligence Public Policy, Microsoft
That's a good question. I would say thank you very much for the invitation to submit materials. We would really appreciate that opportunity. We think the work the committee is doing here is very important, and we want to be as supportive and helpful as possible. I appreciate the opportunity to send some materials, and we will do that.
In terms of the metaverse, everyone is getting very excited about the opportunities there, and I think that is right. There will be a number of technologies that go into creating the metaverse and ensuring that it is performing in a way that people are excited about and is responsible.
I think facial recognition will be one of those technologies, alongside a whole host of other technologies. The metaverse—we call it it the “multiverse” at Microsoft—offers a huge number of opportunities that we're really only just starting to explore as a society. There's a big conversation that we should have around exactly what we want the metaverse to look like and what the safeguards are that we need there to make sure we're reaping the benefits of the technology and addressing some of the challenges.
Facial recognition will definitely be part of that in all kinds of ways that we probably can't even fully appreciate at this point.
4:05 p.m.
Bloc
René Villemure Bloc Trois-Rivières, QC
Thank you very much for your response. This is a conversation we will probably have in another committee, but it is very interesting.
Could you tell me what industries Microsoft's major facial recognition clients are in?
4:05 p.m.
Director, Responsible Artificial Intelligence Public Policy, Microsoft
There is a real mix of different sectors. A lot of the applications we use ourselves. These are things like Windows Hello on our Microsoft devices. If anyone has a Surface or a Windows device, Windows Hello is a big part of that.
More broadly, there are a lot of security and verification applications. Particularly banking and aviation have been exploring this type of security approach. We have banking in Australia, for example, that has been exploring using facial recognition to do PIN-less interactions at an ATM. You would just verify yourself with your face at an ATM to withdraw cash.
Those are a lot of the applications. They tend to be around verification and security, getting into devices and such.
4:05 p.m.
Conservative
4:05 p.m.
NDP
Matthew Green NDP Hamilton Centre, ON
My first questions will be through you, Mr. Chair, to Mr. Larter.
Mr. Larter, I'm going to put some questions to you in a rather rapid-fire way. When you hear me say “thank you” as an intervention, it's to take back my time and move on to the next question. Don't take it as a personal affront, but I have an urgency in the way in which I'm going to ask these questions.
I've heard today that you have not banned the use of FRT technology in Canada with law enforcement. Which agencies in the military, the police and law enforcement do you currently have contracts with, both past and present?
4:05 p.m.
Director, Responsible Artificial Intelligence Public Policy, Microsoft
In Canada?
4:05 p.m.
Director, Responsible Artificial Intelligence Public Policy, Microsoft
I am not aware of any contracts at present, but I do want to make sure that I am getting you an accurate answer to that question, so—
4:05 p.m.
NDP
4:05 p.m.
Director, Responsible Artificial Intelligence Public Policy, Microsoft
I would have to check that. When it comes to facial recognition, I don't believe so, but I would have to go and check that.
4:05 p.m.
NDP
Matthew Green NDP Hamilton Centre, ON
Mr. Larter, just for the record, you're the director of public policy in the Office of Responsible AI at Microsoft. If you were to have contracts with law enforcement locally here in Canada, as the director, would that have gone past your desk? Would you have been made aware of those contracts? Would you have had to authorize and sign off on them?
4:05 p.m.
Director, Responsible Artificial Intelligence Public Policy, Microsoft
Yes, I wouldn't have had to authorize them and sign off on them, but there definitely would have been conversations from across the company, including the wider team that I am a part of, the Office of Responsible AI, which likely would have been involved in that, but—
4:05 p.m.
NDP
Matthew Green NDP Hamilton Centre, ON
Thank you very much.
In your statements, you stated that you did have contracts—i.e., it's not banned in Canada, because I think you stated that you thought we had sufficient legislation in place, a framework that would have been more robust than in the States. I'll give you the opportunity now, as the director of public policy, to perhaps offer with specificity which parts of our privacy and PIPEDA laws you think warrant the use for this, should your company have contracted with law enforcement, the military and the other agencies, as suggested.
