Evidence of meeting #18 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 technology.
A recording is available from Parliament.
12:10 p.m.
Conservative
The Chair Conservative Pat Kelly
Thank you, Mr. Bezan.
Ms. Thompson, thank you for joining us today at the ethics committee. Welcome.
Please go ahead for up to five minutes.
12:10 p.m.
Liberal
Joanne Thompson Liberal St. John's East, NL
Thank you, Mr. Chair. I'm delighted to be here.
My question is for the Privacy Commissioner, Mr. Therrien.
Earlier today, you referenced a moratorium. I believe you indicated that you would support this around the use of facial recognition technology by law enforcement until the regulatory framework around the technology is in place. Would you elaborate on that, if indeed you would be in support of such a moratorium?
12:10 p.m.
Privacy Commissioner of Canada, Office of the Privacy Commissioner of Canada
Our starting point as commissioners—me and my provincial and territorial colleagues—is that the law should ultimately define “allowable” and “prohibited” circumstances for the use of FRT, facial recognition. That's because we are of the view that there are compelling circumstances in which that technology should be usable by police forces. I would not be in favour of a complete ban of the technology, because it does allow use in compelling circumstances.
In my reference to the RCMP, I was suggesting that short of the legislation we truly hope will be adopted in the not too distant future, if the RCMP were to undertake to use that technology only according to a policy—and the RCMP representative last week identified certain characteristics of that policy as being “targeted, time-limited”, etc.—that would be a voluntary partial moratorium, if I may use that expression.
With regard to a complete ban on facial recognition until a new law is adopted, I would not be in favour of such a ban.
12:10 p.m.
Liberal
Joanne Thompson Liberal St. John's East, NL
Thank you.
Would this apply to the use of facial recognition in public spaces as well?
12:10 p.m.
Privacy Commissioner of Canada, Office of the Privacy Commissioner of Canada
Yes, the recommendations that we are making would apply to public spaces as well.
12:10 p.m.
Liberal
Joanne Thompson Liberal St. John's East, NL
I'd like to move on to personal information protection in electronic documents. I have had a lifetime of working in this area.
Currently it's technology neutral, which allows it to endure over time. Should a technology-neutral law apply to facial recognition technology?
12:10 p.m.
Privacy Commissioner of Canada, Office of the Privacy Commissioner of Canada
We're back to the question that I answered a few minutes ago.
Principles-based, technology-neutral legislation for the private sector makes sense as a starting point. The reason we're recommending that there be specific legislation for police forces has to do with the harms of that particular technology of facial recognition. It may well be that certain uses of the technology by private companies raise extremely high risks, not only to privacy but to other rights. Clearview is a good example. We call that mass surveillance.
Mr. Fergus referred to other circumstances. I agree that to denote emotions in order to sell a product, or for whatever other purpose, should not be allowed.
We will provide a few examples of good pieces of legislation. There's draft legislation in the European Union, not yet adopted, which is a good model. Obviously, it would have to be adapted. It says, among other things, that facial recognition should not be used to violate human rights That applies horizontally, whether to the state or to private companies. That is something that I think Canadian parliamentarians should seriously consider.
12:15 p.m.
Liberal
Joanne Thompson Liberal St. John's East, NL
Thank you.
One of the threads that I keep coming back to in the conversation this morning is how we align the realities of the speed of the technology around facial recognition with the need to methodically establish the legislation and the protection around human rights, security and privacy. How do you create that balance?
12:15 p.m.
Privacy Commissioner of Canada, Office of the Privacy Commissioner of Canada
It is through principles-based legislation, augmented—when need be, given the context—by more specific legislation.
I will add this. I heard you ask certain witnesses at this committee if it's too late. It's never too late. Actually, the fact that certain practices are currently occurring should be no reason for you to prevent yourself from doing the right thing and regulating the technology in a way that respects the rights of Canadians.
We are living, not completely but in part, in a world of self-regulation that has led to certain unacceptable practices. It's not because they are routine or banal, as my colleague Diane Poitras would say, that they should continue to be authorized.
12:15 p.m.
Conservative
The Chair Conservative Pat Kelly
Thank you, Commissioner.
In keeping with the day here, were tacking on about an extra 35 or 40 seconds to each person's round.
Go ahead, Mr. Villemure. You're next, for two and half minutes or so.
12:15 p.m.
Bloc
René Villemure Bloc Trois-Rivières, QC
Thank you, Mr. Chair.
A number of studies we've looked at lack conclusive data, but there is nevertheless the possibility of determining things like people's sexual preferences and political opinions, and making such distinctions possible. Are we talking about an unreal world or do we need to look into this matter in the near future, Mr. Therrien?
12:15 p.m.
Privacy Commissioner of Canada, Office of the Privacy Commissioner of Canada
There's nothing unreal about it, and the legislation put forward in Europe, which I just mentioned, is designed to prohibit such practices, because they constitute a genuine risk already.
12:15 p.m.
Bloc
René Villemure Bloc Trois-Rivières, QC
It's a fascinating tool for making distinctions.
We've also heard about biometric terrorism, which is the corruption of databases as people enter and leave the country, to facilitate criminal behaviour. Have you had anything to do with this type of information, not necessarily at the commissioner's office, but in your research generally?
May 2nd, 2022 / 12:15 p.m.
Privacy Commissioner of Canada, Office of the Privacy Commissioner of Canada
Can you give us a little more detail about terrorism? Are you talking about people who might want to enter the country by falsifying data?
12:15 p.m.
Bloc
René Villemure Bloc Trois-Rivières, QC
Would it be possible to facilitate someone's entry by changing all the biometric data so that someone could enter under a false identity? Can the biometric data be falsified at the entry and exit points for criminal purposes?
12:15 p.m.
Privacy Commissioner of Canada, Office of the Privacy Commissioner of Canada
It's not impossible. That gets us back to the idea of protections on the use of facial recognition. Extremely tight security is needed to prevent such risks.
12:15 p.m.
Bloc
René Villemure Bloc Trois-Rivières, QC
All right.
Ms. Poitras, you said earlier that people had not given consent in the Clearview AI case, but all our witnesses have told us that such consent would be impossible to obtain for the use of mass facial recognition. What would you suggest in this regard?
12:15 p.m.
President, Commission d'accès à l'information du Québec
That's a good question, because obtaining consent from people in the context of facial recognition is not always appropriate.
First of all, there is a power asymmetry, whether between the citizen and the state or the citizen and a major corporation, like the web giants.
Secondly, it's difficult to give informed consent, which is one of the essentials of consent. It's an extremely complex technology, and a citizen's ability to give informed consent is, in my view, very limited. The way to mitigate the consent issue consents is to legally authorize some uses, such as some of the recommendations made today. One could also prohibit certain forms of utilization, where it is believed that even with consent or authorization, its use would not be appropriate in a democratic society. I believe that if acceptable and unacceptable ways of using the data were to be set out in the legislation, it would be a step in the right direction.
12:20 p.m.
NDP
Matthew Green NDP Hamilton Centre, ON
Thank you, Mr. Chair.
You will know we have spent quite a bit of time trying to begin to understand facial recognition technology, yet I believe AI offers perhaps an even more expansive way in which public sector and private sector interventions in our day-to-day lives are rapidly shifting our social context. I think about Minority Report. I think about the police's rhetoric around proactive policing and their ability to do predictive policing.
My questions are to the Information and Privacy Commissioner of Ontario, who participated in the process that led to a policy on the use of artificial intelligence technologies by the Toronto Police Services Board by providing comments on the draft policy prior to public consultation.
Were the recommendations you made to improve the draft policy reflected in the final policy?
12:20 p.m.
Commissioner, Office of the Information and Privacy Commissioner of Ontario
Certainly all of our consultations with the Toronto Police Service, including the oversight board, tend to be highly constructive. I point specifically to our consultation, for instance, on body-worn cameras, which resulted in an overarching framework that has since been published.
With respect to the artificial intelligence framework they developed recently, and their policy, we were consulted. We made a number of recommendations, not all of which were adopted, and we continue to consult with them in the development of the procedures.
12:20 p.m.
NDP
Matthew Green NDP Hamilton Centre, ON
Just to be clear on that, which ones of importance would you note today were not adopted by the TPS?
12:20 p.m.
Commissioner, Office of the Information and Privacy Commissioner of Ontario
If I may, Mr. Chair, I would like to ask Vance Lockton. He did the analysis and compared our recommendations with the ultimate policy.
12:20 p.m.
Vance Lockton Senior Technology and Policy Advisor, Office of the Information and Privacy Commissioner of Ontario
Thank you.
I wouldn't say there's anything important that wasn't adopted within the policy that can't be adopted within the procedures.
There was a lot of discussion about getting better definitions of risk levels or a better understanding of how some of the oversight was actually going to happen. We have accepted that it's understandable that this high-level policy may not have it, but it's going to be important to see in the procedures that implement that policy.
12:20 p.m.
NDP
Matthew Green NDP Hamilton Centre, ON
Mr. Chair, I'd like to ask this to all members who are present today. I'm very interested in finding out what their analysis is on artificial intelligence—and maybe they can perhaps provide it in writing to the committee—as it relates to the shifts in ideologically motivated violent extremists and the way in which algorithms and social media are impacting the social context. I reference the recent disruptions here in the nation's capital and other instances across the country.
Thank you.