Evidence of meeting #15 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 used.
A recording is available from Parliament.
12:50 p.m.
Strategic Advisor and Foresight Expert, As an Individual
The only thing I could add to my esteemed panellists here is another potentially benign use case. They may contradict me on this, but one of them is using FRT to prevent industrial accidents. If there are operators who are tired or sleepy—this could include long-haul truckers, others in nuclear or other kinds of industrial facilities, or in health, where they're falling asleep and not being alert to one's lack of attention—it could be potentially beneficial.
12:50 p.m.
Conservative
12:50 p.m.
Computer Science Graduate Researcher, Princeton University, As an Individual
It's not always clear what your image is used for. I don't know that phones these days do this, but they could be collecting that data and using it to turn other models. Consent isn't always clear, and that applies to all of these cases.
12:50 p.m.
Conservative
The Chair Conservative Pat Kelly
Indeed.
Thank you all.
Mr. Fergus, go ahead. You have your hand up.
12:50 p.m.
Liberal
Greg Fergus Liberal Hull—Aylmer, QC
Can I ask a follow-up question to yours? I thought it was a really interesting line of investigation that you were going on, and I want to—
12:50 p.m.
Conservative
12:50 p.m.
Liberal
Greg Fergus Liberal Hull—Aylmer, QC
Following up on Mr. Kelly's question about the individual use and whether or not there's.... Let's say there's consent. When we use the facial recognition technologies to have access to our phones, are those images shared beyond the use of that phone and the owner of that phone? Are fingerprints used? Is that information shared beyond?
I thought the consent on those kinds of phones or security devices was between the end-user and the phone itself. Please correct me if I'm wrong, because I'd like to know.
12:50 p.m.
Prof. Rob Jenkins
Different companies have taken different positions on that. For some companies, everything happens on the device. For other companies, that's not a part of the deal; it can go to the cloud and from there to other places.
12:50 p.m.
Conservative
The Chair Conservative Pat Kelly
Would any of our other witnesses care to comment?
Go ahead, Dr. Watkins.
12:50 p.m.
Postdoctoral Research Associate, Princeton University, As an Individual
Thank you.
I agree with Dr. Jenkins. The variance that he described, the lack of certainty that we have about whether the data stays on the phone or if it stays on company servers, or if, in fact, it's used by a third party vendor and stored on their servers, shows the need for transparency that we ought to have on where this data is being stored and how it's being used.
12:50 p.m.
Computer Science Graduate Researcher, Princeton University, As an Individual
There's a new version of training called “federated learning”, where you can keep the image on your device the entire time, but you still consent to an update. You tell them all how it should adjust its parameters such that it can better classify your own image. In this case, the image never leaves you or your phone, but the model is still able to use that information to improve itself. However, it is a bit ambiguous how consent would work there.
12:50 p.m.
Conservative
The Chair Conservative Pat Kelly
All right.
Thank you so much to our witnesses. It's been a very informative panel.
With that, the meeting is adjourned.