Evidence of meeting #116 for Access to Information, Privacy and Ethics in the 42nd Parliament, 1st Session. (The original version is on Parliament’s site, as are the minutes.) The winning word was advertising.

A recording is available from Parliament.

On the agenda

MPs speaking

Also speaking

Taylor Owen  Assistant Professor, Digital Media and Global Affairs, University of British Columbia, As an Individual
Fenwick McKelvey  Associate Professor, Communication Studies, Concordia University, As an Individual
Ben Scott  Director, Policy and Advocacy, Omidyar Network

11:55 a.m.

Director, Policy and Advocacy, Omidyar Network

Dr. Ben Scott

I was going there next; you beat me to it.

When Facebook, Google, and Twitter announced they were going to do ad transparency databases, they said you're going to see all the ads that are run during a political cycle and you're going to have data about every single one of them: who bought the ad, how much they spent, and some information about targeting, although they've reneged a little on that. It was all going to be searchable and there was going to be API access so that researchers and journalists could literally download the entire data set and study it themselves.

That last piece has not been done. The searchability and the research capability of that data set are not up to the standard we need. I don't believe the companies are going to get there on their own, nor should we expect them to. These are businesses. They don't want to reduce the amount of commercial advertising revenue coming in the front door. Their responsibility isn't to protect the public interest. That's ours.

If we want that standard to be in the market, we're going to have to put it in the law.

11:55 a.m.

Liberal

Nathaniel Erskine-Smith Liberal Beaches—East York, ON

When it comes to subjecting political activities to a privacy regime, it's one of our recommendations, but when we did have Christopher Wylie here, he indicated there ought to be a difference between how we treat political activities and how we treat profit-seeking enterprise, because of course the pursuit of democracy is different from the pursuit of dollars. Do you see there would be a differentiation between the stringency of that privacy regime, or should we apply the same regime to both enterprises?

11:55 a.m.

Director, Policy and Advocacy, Omidyar Network

Dr. Ben Scott

I've wrestled with this question myself. I think you can differentiate them. I think there is a logic to differentiation to suggest that we have a standard for political advertising that is different from selling soap or bicycles, just as we do in broadcast ads.

It becomes more challenging when you start asking what a political ad is. It's interesting to see how the three big platform companies have defined political ads in what they put in their databases and what they apply their voluntary regulation to.

Twitter says an ad is only a political ad if it mentions a political candidate or a party. That's the Google standard as well. Facebook says a political ad is anything that mentions an issue of public importance, and they list about 20, including everything from climate to gun control to immigration. To me that's a much more honest presentation of what a political ad looks like. I think it's a mistake to limit your terms too narrowly, because people will just go around you and use different things as proxies, but once you begin to define it more broadly, the grey zone between what's political and what's non-political becomes more challenging to define.

I think it's like night and day: 95% of the time we can agree whether it's night or whether it's day. We'll adjudicate those 5% of cases that have legitimate opinions on both sides. I think they can be divided. I think that's the right starting point. If we find that a company can't distinguish between the two, fine; you can just make the same high bar for everything.

Noon

Prof. Taylor Owen

Very briefly, I think Ben Scott and I agree on most of this range of policy responses here, but on this one there might be a bit of distance between us. In the long run I'm not sure we can make that distinction between political and non-political ads in a viable and sustainable way.

Ultimately, whether it's for commercial or political activity, both are seeking to influence our behaviour. Why wouldn't we, as either consumers in a consumer protection realm or voters in an election integrity realm, be allowed to know how our behaviour is being microtargeted using incredibly sophisticated systems to target and nudge our behaviour? I see different baselines for both, but the easier solution is to make it all transparent.

Noon

Liberal

Nathaniel Erskine-Smith Liberal Beaches—East York, ON

We talked about advertising transparency. Mr. McKelvey, you talked about collection and use and more stringent requirements on the collection and use of personal information and subjecting political activities of political parties to stronger requirements perhaps. If I knock on a door and I speak to someone and they say they're really interested in climate change and then I target an ad to that individual among other individuals who are all concerned about climate change, do we have a problem with that?

Noon

Prof. Fenwick McKelvey

I think, in principle, getting to your question, political parties can easily fit within our existing privacy law. If you're collecting information about their views at the door, then it seems to me—not being a privacy lawyer—that there's an informed consent. You're asking them for their views, and that's something you're collecting and they know you're collecting it. Then you're using it.

I think it comes into a question, which I think is a question writ large in our data in this kind of combination of surveillance and targeting: When do we know, and when are we informed that information you're collecting is going to be used for targeting purposes?

The point that I'm trying to make is that I'm not convinced that all this targeting is super-effective. If you're a political party and you want to target people about climate change, why do you need to link that to the voter whose door you're knocking on in the first place? I think that there are ways you can abide by the privacy law and still be able to conduct relatively the same type of business.

It gets into whether you're sending a specific targeted email to that person who's talking about climate change. You know, maybe that might be restrained in some way.

Noon

Liberal

Nathaniel Erskine-Smith Liberal Beaches—East York, ON

Do I have much time?

Noon

Conservative

The Chair Conservative Bob Zimmer

You have one second. Unless you can be as quick as Mr. Masse, it's—

Noon

Liberal

Nathaniel Erskine-Smith Liberal Beaches—East York, ON

Maybe you'll get back to me.

Noon

Conservative

The Chair Conservative Bob Zimmer

Okay.

We have five minutes for Monsieur Gourde.

Noon

Conservative

Jacques Gourde Conservative Lévis—Lotbinière, QC

Thank you, Mr. Chair.

I thank the witnesses for being here this morning.

From your discussions, we all understand that the current digital world is evolving, that it is changing and transforming. The various digital platforms provide us with a whole new landscape of information, but also of misinformation, which is unfortunate.

We have been seeing over the past two or three years just how much fake news is taking over all digital platforms. We are wondering whether Canadians will drop out at some point because no one can any longer be sure how accurate the information provided is. Is it real or fake? That is a current issue in the United States, where news is provided one day, refuted the next, but again presented as real the day after.

The public is confused by these new media, which have unfortunately taken over part of the traditional media's market. I think that traditional media were more ethical because they spent more time on their research before providing information to the public.

There is also the issue of advertising on those platforms. I think that new media care little about the accuracy of the advertising they air. Who do you think should be responsible for the veracity of advertising proposed by both the private sector and political parties? You can each answer in turn.

12:05 p.m.

Prof. Fenwick McKelvey

If I understand correctly, the credibility of advertising is a real issue. That's where I've noted companies like Facebook and Google exiting from certain ad markets or restricting ad options for certain keywords, such as “cryptocurrency” and “opioid addiction treatment centre”.

I think it's very important to recognize that this is a clear limitation when you have some of these algorithmic markets being constituted. It's that they suffer at times from being able to recognize quality or credible information.

I think that is one of the ubiquitous problems that we have. To operate at scale, information triage has kind of taken on a market-like approach. I think that it has often failed in delivering high-quality ads, or at least with opioid addiction treatment centres, ads that aren't from scammers or dubious treatment centres.

I think that's an important finding. It's important to recognize that there are real challenges in how algorithmic recommendation takes place and whether that functions effectively in being able to discern what's good and bad information, to use those terms super-loosely.

12:05 p.m.

Conservative

Jacques Gourde Conservative Lévis—Lotbinière, QC

Go ahead.

12:05 p.m.

Conservative

The Chair Conservative Bob Zimmer

Hold on, Mr. Owen. We're not getting....

Okay, go ahead.

12:05 p.m.

Prof. Taylor Owen

I have a quick point on trust of information. I think it's pretty clear that trustworthy information that is known by a large number of citizens is critical to a democracy. We have to have some baseline of trustworthy information on which we are making democratic decisions about our collective well-being and governance. This is critical to a democracy, and that is being eroded by the system.

If we take that as our baseline, then I think we need to look at how we create more trust and more reliable information in the ecosystem we now have in our digital public sphere. Certainly advertising credibility is a part of it, ad transparency, but a big piece is the amount of journalism that is being produced in our society about our society and is holding power accountable within our society, and that is in steep and precipitous decline in Canada.

There are a host of other regulatory changes or points of governance engagement that could help make that more robust. There are easy things, such as changing the Income Tax Act to allow for charitable funding of news. In the U.S., the most robust sector in the journalism space, particularly the accountability journalism space, is non-profit news. This is almost non-existent in Canada because of our charitable funding law.

I think there is a whole host of things we could do, at the very least, to build up that backstop of reliable journalism in this space as well.

12:05 p.m.

Prof. Fenwick McKelvey

I would add to that also, thinking about public broadcasting. I think that one of the ways that we're seeing this issue is that we think about information subsidies, or what's subsidizing the production of information, and I think there is a whole host of new information subsidies. This is when you talk about native content or sponsored content, as well as propaganda campaigns. That's really where, to me, it's also looking to public broadcasting as another important source and realizing that part of the integrity of our democracy is funding public broadcasting.

12:05 p.m.

Conservative

The Chair Conservative Bob Zimmer

Thank you, Monsieur Gourde.

Next up, for five minutes, is Monsieur Picard.

12:05 p.m.

Liberal

Michel Picard Liberal Montarville, QC

Thank you.

In order for the government to regulate, we have to identify the real problem or problems. Let's go back to the basics of the question at stake, the breach of personal information. What is the problem with Facebook and Cambridge Analytica? Is it the fact that someone was intelligent enough to draw conclusions about the behaviour of people based on public information provided by subscribers, or because they did it without our knowing it?

12:05 p.m.

Director, Policy and Advocacy, Omidyar Network

Dr. Ben Scott

To me, the problem is both. I'll answer—

12:05 p.m.

Conservative

The Chair Conservative Bob Zimmer

I'll start with Mr. McKelvey and then go to Mr. Scott.

12:05 p.m.

Director, Policy and Advocacy, Omidyar Network

Dr. Ben Scott

Chair, I have something I want to say on this.

12:05 p.m.

Conservative

The Chair Conservative Bob Zimmer

Sorry, Mr. Scott. We had Mr. McKelvey first, and then we'll go to you.

12:05 p.m.

Director, Policy and Advocacy, Omidyar Network

Dr. Ben Scott

I apologize.

12:05 p.m.

Prof. Fenwick McKelvey

I was just going to say that what is clear is that what's been exposed—and I think what Facebook has also admitted before this committee—is that they have been entrusted with a lot of personal information and data and they have not been discerning about who has access to that personal information. I'm quite skeptical of whether Cambridge Analytica was effective, and I'm not particularly convinced about the psychometrics as some sort of revolutionary new hypodermic needle, but I am thinking that it is very clear that if you're collecting large amounts of data, there is an obligation under the privacy law to make sure that you're controlling the flow and who gets access to it.

I think it's been very clear that this has been one of the key issues here, the kinds of data-sharing arrangements that have taken place in social media.

12:10 p.m.

Liberal

Michel Picard Liberal Montarville, QC

I'll get back to Mr. Scott.

Is it access, or is it knowing what they would do with the data?