Evidence of meeting #25 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 data.
A recording is available from Parliament.
5:20 p.m.
Bloc
René Villemure Bloc Trois-Rivières, QC
Could you give us something in writing on worthwhile policies in the European Union's General Data Protection Regulation, GDPR? We could learn from it.
5:20 p.m.
NDP
Matthew Green NDP Hamilton Centre, ON
Thank you very much, Mr. Chair.
For my own edification, this is a question I'm always wondering about when we have expert witnesses.
Mr. Maslej, can you state whether Stanford and the human-centred artificial intelligence centre are funded by any AI companies? Are there any potential conflicts to put on the record?
5:20 p.m.
Research Associate, Institute for Human-Centered Artificial Intelligence, Stanford University, As an Individual
I'm not familiar with the financial situation that surrounds the institute, but I also want to clarify that I'm speaking here as an individual and, mostly, to present the data that the index has on facial recognition. My views are not those of Stanford.
5:20 p.m.
NDP
Matthew Green NDP Hamilton Centre, ON
That's good. Okay.
The AI Index report indicates that language models are now more capable than ever before, but also more biased. Can you elaborate on that statement?
Why are these models becoming more biased as they become more advanced, and what risk does that pose?
5:20 p.m.
Research Associate, Institute for Human-Centered Artificial Intelligence, Stanford University, As an Individual
Part of the reason they're becoming more biased is that, typically, these models are being fed increasingly large numbers of data. For certain models, it is advantageous to have a lot of data. The more data you give the model, the more likely it is to get some data that is not ideal.
We saw this in the report with this model clip, which is a multimodal linguistic model. This model was asked to assign the probability of an American astronaut, Eileen Collins, being.... The model was asked, “What is this image of?” The model assigned a higher probability that this photograph was of a smiling housewife in an orange jumpsuit with the American flag than that it was of an astronaut with the American flag.
That's not our finding. That's a finding from a paper of Birhane et al., 2021. It's illustrative of the fact that when you give these data a lot of models, which might be required for higher performance, they might catch some conspiratorial and biased data. If we're not filtering that data proactively, it could be very likely that these models behave in toxic and problematic ways.
5:20 p.m.
Research Associate, Institute for Human-Centered Artificial Intelligence, Stanford University, As an Individual
One of the things I allude to—and we talk about this in the report—is the issue of filtration, whereby you can mandate that companies filter the kind of data they use to train their systems. It's been reported in different papers that there can be a filtration tax. That is, if you filter data before you apply it to a model, the model might not perform as optimally as if you gave it unfiltered data, because the more data a model has, typically, the better it can perform on certain tasks.
Filtration can be an avenue to do that, but it also might present some trade-offs for businesses.
5:20 p.m.
Conservative
The Chair Conservative Pat Kelly
We're over again. I'm sorry.
We're going to end up going a little over time here. I promised our last two members four minutes each, so I'll stick with that and go to Mr. Kurek for four minutes.
5:20 p.m.
Conservative
Damien Kurek Conservative Battle River—Crowfoot, AB
Thank you very much, Mr. Chair, and thank you to our witnesses for their expertise and for coming before us today.
Ms. Polsky, with regard to recommendation 19 in your report, if a foreign national is negatively impacted by an in-Canada FRT project, what would you suggest is the action, corrective or otherwise, that should be taken?
5:25 p.m.
President, Privacy and Access Council of Canada
Forgive me, because I don't have the report on screen, but if it affects a foreign national in the same way, I'd say that immigrants to Canada, who are not yet landed immigrants or citizens, have constitutional rights and charter protection, and perhaps that needs to be afforded to them as well. That is something that needs further exploration for sure.
5:25 p.m.
Conservative
Damien Kurek Conservative Battle River—Crowfoot, AB
Thank you for that.
I want to talk about the second recommendation. I appreciate that it's very helpful, just as a note for witnesses, for recommendations to come forward, because that certainly helps committees in the structure of the reports we are able to put together.
On the second recommendation, when it comes to the broad, society-wide database, how could such a database cause harm and possibly result in abuse of those who may be found in that database? The inverse of that is, for those who are not in that database, how could they possibly be harmed as well?
5:25 p.m.
President, Privacy and Access Council of Canada
A simple example of something that's going on in Canada, unregulated once again.... We surveyed the residential tenancy boards across the country. It's an app that if you want to rent an apartment or a house, it's a tenancy app. You can't put pen to paper and fill out an application form. You must use this app. It also creates, if you will, a blacklist of tenants, because the landlords can put in any information or comment, like, “She was late on her rent by two days,” or, “She has a kid that's loud.” They can put in whatever they want, and other potential landlords can look at this and say, “I'm not renting to this person.”
What happens to the person who wants to rent a home and doesn't know this comment exists? They have no recourse. Do they become homeless, as a result? In the United States, apparently, this is going on, and some of these also require that the applicants submit their facial biometrics and other biometrics, including very personal information that wouldn't be allowed to be requested elsewhere. These have profound implications.
5:25 p.m.
Conservative
Damien Kurek Conservative Battle River—Crowfoot, AB
It's interesting, because recently I was watching a television show, and that was one of the case studies with profound possible impacts.
I'd like to move from the tenancy example, if I could. There's been a collection of data at Toronto Pearson, the international airport that I'm sure all of us on this call have been through countless times. It includes, I can only imagine, an unbelievable amount of data about Canadians, individuals visiting our country, and everyone in between. I'm curious if there are any further thoughts, recommendations, or concerns that you would highlight, and how an organization like an airport, or another entity, with law enforcement—
5:25 p.m.
Conservative
The Chair Conservative Pat Kelly
You've used up all of your time. We don't have any time for an answer. I'll allow Ms. Polsky maybe five seconds, if she has a very succinct point to make in response to that lengthy question.
5:25 p.m.
President, Privacy and Access Council of Canada
Between Beyond the Border, Preclearance, the Customs Act, and all the legislation that has an impact on this, we have to look at that not in isolation but as a whole for the entire system.
5:25 p.m.
Conservative
5:25 p.m.
Liberal
Lisa Hepfner Liberal Hamilton Mountain, ON
Thank you, Chair, and thank you to the witnesses for your testimony today.
Mr. Maslej, you talked about this in your statement, but I found it really interesting that in the research from the 2022 AI Index report, facial recognition technology in 2017 had a 50% error rate, and by 2021 there were no platforms with an error rate greater than 3%.
Can you reflect on that? Why have we seen such technological advancement? How has this happened? What are the ramifications and implications for the future based on that?
5:30 p.m.
Research Associate, Institute for Human-Centered Artificial Intelligence, Stanford University, As an Individual
Yes, that's a great question.
I would say that the reason you have seen this technological advancement is that AI systems are getting better across the board. There have been a lot of developments in the architecture that powers these systems. These systems now run on better hardware, which means they're able to operate at much faster speeds. There are a lot of academic and business reasons that these systems are operating better.
In terms of consequences, again, all of this portends and points to the fact that AI is going to become a part of our lives whether we like it or not and, as I have said to you today, there is a lot that AI systems can do, but there's also a lot they can do that we didn't expect them to do or didn't perhaps want them to do. Rather than just welcoming these systems into our lives with open arms, it is important to ask what kinds of effects they might ultimately have.
Again, if we live in a world where, for instance, FRTs are now having incredibly high success rates, it might be a lot easier for companies to justify that they ought to be used, but again, that doesn't necessarily imply that we shouldn't think critically about how they ought to perhaps be regulated or managed by government officials.
5:30 p.m.
Liberal
Lisa Hepfner Liberal Hamilton Mountain, ON
Thank you very much. That is a very helpful answer.
Ms. Polsky, you and I have spoken already, and you've spoken a bit about this today, but you make a really good case for the fact that we don't have enough education around facial recognition technology. I'm wondering what your suggestions are around education. How do we improve this? What do we need to do specifically around education?
June 9th, 2022 / 5:30 p.m.
President, Privacy and Access Council of Canada
I think it will be important to mandate that for the Privacy Commissioner—and this goes for the individual provinces as well that have substantially similar legislation—all legislation require that, first, the privacy commissioners' offices be fully funded and that they have a separate fund, also fully funded, for education. They don't all have an education mandate. They do some work to educate, but there needs to be a much more formalized program, because that will translate into people being more aware of the legislation and their rights and responsibilities.
They can build that into the technology, and then, once they have the technology, one thing that could be done is to test it in a sandbox, a neutral sandbox run by the Privacy Commissioner, as an opportunity not only for the commissioners and civil society groups to examine it, but for the corporations to allow it to be examined in a trusted neutral setting that does not violate their copyright or their intellectual property. That way, it gets tested and approved before it's allowed for sale in Canada.
Also, fund the education through the Privacy Commissioner's office, and again, without the influence of industry, please.
5:30 p.m.
Liberal
Lisa Hepfner Liberal Hamilton Mountain, ON
Thank you very much.
I have only a few seconds left, so I'll just thank you all for your time today. It was very helpful and very interesting.
5:30 p.m.
Conservative
The Chair Conservative Pat Kelly
Thank you.
With that, I thank our witnesses, although I must now ask them to disconnect as quickly as possible. We will be disconnecting this Zoom and beginning an in camera Zoom for, hopefully, just a couple of quick minutes of committee housekeeping business.
With that, the meeting is suspended pending the in camera Zoom call.
[Proceedings continue in camera]