Evidence of meeting #118 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 content.
A recording is available from Parliament.
11:45 a.m.
Assistant Professor, Department of Communication, University of Ottawa, As an Individual
Thank you.
I also haven't read the details of USMCA yet, but I think the questions that are brought up are important and we need to look into them.
To build off what Ms. Bradshaw just said, I would say that the idea of platforms being responsible for all of the content that shows up on them has been a major question. As she said, it has been a really valuable tool for the growth of these companies and for innovation in how we deal with the mountains of data and information that now exist, and that's helpful. However, I think that there's an important distinction to be made between allowing content to exist and being responsible for that content, and being responsible for what content shows up as trending topics, recommended search results or something that is at the top of people's newsfeeds.
There are decisions that these platforms make already about what gets the light of day and what doesn't, and those decisions need to be considered in terms of whether or not they are silencing groups that shouldn't be silenced or promoting racist or hateful content that shouldn't be promoted and put forward.
Without having looked at USMCA, I can't tell whether or not that distinction between content and the dissemination of content has been addressed, or the implications, but I think it's an important thing to look at.
11:45 a.m.
Conservative
Peter Kent Conservative Thornhill, ON
Some of the early conversation has seemed to suggest that there's a possibility that it would place North America and the protection of privacy in North America at odds with the GDPR, for example, so that the North American regime versus Europe, certainly, and perhaps other jurisdictions, would be at odds.
11:45 a.m.
Assistant Professor, Department of Communication, University of Ottawa, As an Individual
It seems like it will cost a whole lot of companies a whole lot of money and will be problematic for citizens if that is the case, but I don't know the specific details.
11:45 a.m.
Conservative
Peter Kent Conservative Thornhill, ON
Mr. Pal, could I just ask about your thoughts on the GDPR, which came into effect less than six months ago now, but has already affected the behaviour of some North American—particularly American—news organizations by causing them to cut themselves off from distribution in Europe because of fear of prosecution?
11:45 a.m.
Prof. Michael Pal
I think there are some things that we can learn from the GDPR. I'm not even sure that European privacy experts have a really concrete understanding of how it's really going to operate in practice. What does it mean for news organizations? What does it mean for technology companies?
I think what it does point to, though, is that a lot of these tech companies are based in the United States or are based internationally, so in some sense, each domestic state has lost control over some of the conversation around elections. It's not just CTV, CBC and the other domestic broadcasters. I think that it's incumbent on each of the states to try to....
We shouldn't rely on the GDPR, on laws in California or on U.S. federal law. We, as Canadians, have to come up with our own set of rules. I think Professor Bradshaw made the important point that we want to facilitate political expression. We don't want to restrict that. Some of the potential laws you could come up with might restrict political expression. It's a charter-protected right—paragraph 2(b)—and we don't want to restrict that.
Where the charter considerations are lessened, though, is in the area of foreign interference, or foreign actors, individuals or entities expressing views on Canadian politics. Regulating Internet companies or voter privacy in a way that restricts foreign interference, I think, stands on much firmer theoretical and legal constitutional statutory grounds.
11:50 a.m.
Prof. Michael Pal
I'm not critiquing Ms. Bradshaw. I think she made a good point on that, but that's the broader constitutional implication.
11:50 a.m.
Conservative
The Chair Conservative Bob Zimmer
Yes, you bet.
Thank you, Mr. Kent.
Next up for seven minutes is Mr. Angus.
11:50 a.m.
NDP
Charlie Angus NDP Timmins—James Bay, ON
Thank you very much for this.
I want to talk about the government's decision not to put the political parties under PIPEDA or a similar regime that would respond to the specifics of the differences between political parties and commercial entities. If you talk to the people around our committee, we're very concerned about this, because we have been looking at the ability of third party actors to actually undermine democracies. If you talk to our political organizers, they're over the moon, because we are now in a digital arms race, and no political party is going to be willing to put down their arms first because they see the potential to do more and more targeting, to get more and more precise, and to shift votes in areas that are key. That's the reality we're dealing with, and we're dealing with a completely different world than we were in 2015 in terms of the speed with which this is happening.
Mr. Pal, I would like to ask you about the importance of having some kinds of provisions in place to make sure that we don't end up misfiring with these weapons, because we've seen what happened in the U.S. and we've seen with the Brexit vote how it can be manipulated. Do you think we need a strong legislative regime, and what would it look like to have some manner of accountability with regard to how our parties and other actors use this information politically?
11:50 a.m.
Prof. Michael Pal
You make a very good point that the scenario in 2019 is very different from what it was in 2015. Things move very quickly, and the risk is that you could put in regulation that is overly intrusive, or that doesn't actually achieve what you want or has the wrong consequences. I'm always aware of that, but I think we have a lot of good information. Political parties are collecting enormous amounts of data, personal data, sensitive data. Parties have always done so, but it's just reached another level. To me this is a non-partisan issue. It doesn't affect one party more than the other, but currently all political actors have an incentive to up their data operations and their data game.
The pitch that I try to make in this space is that actually privacy rules are to the benefit of political parties. No one wants to be regulated, and it may seem onerous and it may cost money, but imagine what would happen if there was a hack of one of Canada's major political parties, similar to what happened with the Democratic Party in the United States. It wouldn't take many hacks, or many instances of personal information being disclosed by, say, a malicious foreign actor for the public to potentially lose faith or trust in that political party or the system as a whole. I think we are at a moment where it's very important to address the privacy issues, and doing so is in the interest of the political parties themselves.
I tried to suggest a few areas in terms of content, such as the right to know what data a political party holds about you, and the right to correct incorrect information. A lot of hard work is done by volunteers, as you all know, and when you're entering information on an app or on a piece of paper, it's very possible for information to be incorrect, and that may be something the voter, the individual, doesn't want. I think rules on who gets access to political party databases or at least disclosure about that might be helpful as well.
I understand those may at times seem onerous to political parties, but I think they go a long way to instilling confidence in voters that the parties have their best interests in mind. The worst case scenario is a hack. We've seen denial-of-service attacks on political parties. I believe the Prime Minister summarized the Communications Security Establishment report, which said there were low-level attacks in the 2015 election. CSE said there were over 40 incidents of interference around the world, so we shouldn't see Canada as isolated from that. I have a lot of concerns about 2019, and I think privacy addresses some of those.
11:55 a.m.
NDP
Charlie Angus NDP Timmins—James Bay, ON
Ms. Bradshaw, I'm very interested in your analysis of junk news. You talk about “cyber troops”, psy-ops, the weaponization of AI.
In 2015, in my region, because I basically live on Facebook, according to my wife, I saw a completely different narrative than what was in the national media. I saw deeply racist posts, mostly from Britain First, anti-Muslim posts, posts attacking immigrants and attacking refugees. They were very targeted. They were targeting on Facebook in key areas of my region among key voters.
It was a completely different message than anything that was happening nationally. It wasn't really noticed, because we still pay attention to what Peter Mansbridge says at six o'clock.
I always felt that out of that there had to have been a better or clearer type of targeting, such that these Facebook users who were not normally political were suddenly repeating this type of message. This seemed to be what we saw out of Brexit, the idea that groups such as Cambridge Analytica can specifically target the poll voters, the voters who are actually going to be influencing, and going on and pushing this.
You talked about how this was used in swing states with certain swing voters. I'd like you to elaborate on that.
I'd also like you to elaborate...because we keep talking about the third party actors as though it's just the bad, hired mercenary guns. You talked about the political influencers who actually are in the parties. Can you talk about the connections between people in the parties, these third party operatives, and how they're using this misinformation online?
11:55 a.m.
Researcher, As an Individual
If we look at the U.S. and junk news being spread in swing states, this is just based on Twitter. It wasn't a Facebook analysis, but just Twitter and what people were sharing as news and information.
We analyzed a couple of million tweets in the 11 days leading up to the vote. If you looked on average at the URLs that users were sharing in swing states, they tended to point to higher rates of junk news and information, compared with uncontested states. Therefore, part of this is the somewhat organic drive of spreading misinformation. It's not necessarily coming through the advertisements but it's being organically spread through the platforms by users, or maybe by bots, who did play somewhat of a role in amplifying a lot of those stories.
The way we measured where the accounts were coming from was by using geo-tagged data. If a user had reported to be in Michigan, for example, which was one of the swing states, that's how we determined where the information was and where the junk news was concentrated.
There's the organic side of it, but there's also the targeted advertisement side of things. We have a lot of information on Russia, thanks to Facebook's disclosures around Russian operatives buying political advertisements and targeting them to voters based on their identities or values. They homed in on groups such as gun-right activists and the Black Lives Matter movement.
They tended to also play both sides of the political spectrum. It wasn't only about supporting Trump. They also supported candidates such as Jill Stein and Bernie Sanders. They never supported Clinton, though. They would always launch ad attacks on her.
The stuff that comes from the political parties themselves is really hard to trace. That relates back to the question you asked before on laws and what we can do to improve some of this targeting stuff.
We talked to and interviewed a lot of the bot developers who worked on campaigns for various parties. They were the ones who created the political bots to amplify certain messages. It's hard to trace their work back to a political party because of the campaign finance laws that only require reporting up to two levels. Generally how these contracts go out is that there will be a big contract to a big strategic communications firm, which will then outsource to maybe a specialized Facebook firm, which will then outsource work to a bunch of independent contractors. As you go down the list, you eventually get to the bot developer, who we interviewed.
We don't have any specific data on exactly what parties these groups worked for, at least none that I can share because of our ethics agreements with these developers. The big problem here is that we're unable to actually track because of campaign finance laws.
Noon
Liberal
Michel Picard Liberal Montarville, QC
I need an introduction that may be long, so bear with me.
In my comments, I will disregard and not consider disinformation like calls that send someone to the wrong poll or anything covered by the Criminal Code.
In talking about junk news, fake news, whatever, in the Cold War, and especially in war times, propaganda was one of the best tools in town to make sure that your message, whatever it was, went through. This, and magazines, photos in Middle East countries, in which you see food, everyone at the table, big cars, well-dressed people, just to push the population against their own government....
At the time, sending a thousand letters for publicity, whatever it was, cost a fortune. You had to send one or two pages. Today you send five million to 10 million emails in a click—no cost.
I will submit for your consideration that you are looking at the problem from the wrong end. That's my hypothesis. We try to focus on those who provide this information and bad content, and try to regulate company's social media because they do things that are not good, probably because people are too lazy to do their own cross-checking and verification of information. By the way, we don't prevent people from seeing specific information. We just download a huge amount of information and you don't see where you are anymore.
From a regulatory standpoint, how do you expect me as a government to act on those companies that are sending this kind of content without touching their freedom of speech?
Noon
Assistant Professor, Department of Communication, University of Ottawa, As an Individual
I can start.
I understand that you want to separate things like voter suppression tactics from junk news and junk content. However, when we're thinking about how to deal with one of those things and not the other, it's very difficult to say that we would regulate the platforms only for one thing and we're going to have a completely different solution for the other. The conversations go together, because the mechanisms for getting the information to the front page of somebody's newsfeed are the same.
There's that sort of technical challenge there, but then I think the idea of how we balance this against questions about free speech is a really important one. We don't want to have a democracy where there are people who don't get to share their opinions, where certain views are silenced. That is certainly a problem.
We have to think about the changing media system, though. We have to think about the fact that it used to cost a lot more money and take a lot more resources to spread disinformation. Now it's very easy to spread it.
We also know that people used to not have a whole lot of choices in terms of what content they were getting. For them to be media literate, and not be lazy, in your terms, was a lot simpler. There were fewer checks that they needed to do. There was less work that they had to do to make sure they knew what content was showing up and who created it. There were only so many people who could afford a broadcast licence.
The expectations we put on the citizens in that context are very different from the expectations we would put on citizens now, in saying, “Look, we can't regulate platforms. This is the responsibility of citizens.” In the media environment that we have, I think it's unreasonable to expect citizens to be able to discern the different sources of content, what is true and what isn't, without some support.
I don't want to let citizens off the hook. I think that digital literacy and media literacy are very important, but I think it's one piece of a larger puzzle.
Noon
Researcher, As an Individual
A lot of it comes down to addressing some of these bigger systemic issues around the platforms. If you can address the root cause of bad information and junk information being able to spread so quickly across social media platforms, then you could also address that while protecting freedom of speech.
Many of the systemic problems, to me, have to do with the whole idea of the attention economy and how, in a world where information is everywhere, attention becomes our scarce resource. Platforms are built on that attention economy. They're designed to tailor content and information and advertisements to us that are going to draw us in, which is what a lot of this junk news does. It's clickbait content. It's designed to get our emotions going and to make us feel angry or happy, to get our attention.
If you can start addressing the way social media tailors content to users by looking at the actual principles that go into their algorithmic design—relevance and virality, for example, are things that are very important to the algorithm right now—and switch those to be principles that would support better democracy, then you can start to regulate the platforms in ways that wouldn't harm free speech.
12:05 p.m.
Prof. Michael Pal
Regulating fake news is the hardest issue. My fake news might not be your fake news. Voter suppression's already illegal. Foreign interference is already illegal, and C-76 takes some really good steps toward closing the final loopholes that are there. C-76 would put in place an offence of impersonating a politician or a political party, so that you couldn't purport that the advertisement came from a particular elected representative or party.
In the Canada Elections Act it's already illegal, if you're a candidate, to claim the person you're running against is going to drop out of the race for some reason. That was a common tactic that was used, so we legislated against it. In C-76, voter suppression and robocalls are all things where we've updated the legislation to deal with whatever the new dirty tricks are, potentially. I don't see any problem in saying, “Today, a lot of the dirty tricks are potentially happening on social media. Which entities have the resources and the ability to actually ensure that the rules are followed?”
It's impossible to try to track down every purveyor of misinformation, disinformation and voter suppression. It's much easier to regulate the social media platforms. It's technologically feasible. They're telling us they're changing the world. They should be able to have a transparency and a repository of election ads without too much of a hit to their bottom lines, and I think we can do that in a way that respects freedom of political expression. One of the things we've learned over the last 18 months is that regulating social media platforms is actually what has to happen now.
12:05 p.m.
Conservative
October 2nd, 2018 / 12:05 p.m.
Conservative
Jacques Gourde Conservative Lévis—Lotbinière, QC
Thank you, Mr. Chair.
I thank the witnesses for being here this morning.
I'd like to go back to the roots of the communication problem.
In an election, candidates have to reach between 90,000 and 110,000 electors in a short period of time, in approximately two months. We, the members, have more time because we are already in our riding. If we want to be fair to everyone, that period lasts a maximum of two months.
It is a big challenge, because it's difficult to reach people. Elections Canada provides us with an address and a name, period.
There are two ways to reach the electors, and that is to go door to door or use their phone number, but there are fewer and fewer landlines. All we can do next is try to find cell phone numbers, which is more or less legal, because those are considered confidential in Canada.
Using digital platforms has become a necessary evil for all future politicians if they want to reach a large number of people in a very short period of time. We have less than a 1,000 hours to reach 100,000 people, and that may not add up to a lot of minutes per person. That could be why targeting becomes very interesting to politicians.
Do you think it would be possible to create legislation for these platforms, so as to restrict access to some information, or give politicians more access to these platforms in order to reach people?
If we can't get the phone numbers, we turn to digital platforms. If politicians had access to cell numbers, they could call people and speak to them to at least have some direct contact.
Currently, we are doing politics through indirect contacts. To reach an elector, we are using machines and artificial intelligence, and that is not the essence of democracy. We aren't electing robots, but members of political parties and a prime minister.
The time we have to reach electors is very limited, and I see a problem there for democracy in the short, medium and long term, and that is the root of the problem. Everything else is related to the lack of time and information.
Should Elections Canada provide us with landline telephone numbers and cell phone numbers?
12:10 p.m.
Prof. Michael Pal
One solution might be that political parties should have more consistent information. As you said, the Elections Act provides basic information about name and address. One option would be to provide other information.
You suggested phone numbers. It makes me a little bit uncomfortable.
I don't think people would like political parties to have their cell numbers.
I would have to think about that more. If political parties had the right to have greater disclosure about information about voters, then it would be incumbent to have greater privacy protections for that information. If that's a policy direction that the committee is considering recommending, or that you are as an individual, then there has to be a balance there between increased privacy protections and the information that parties have.
Could there be an update on the basic information that's provided? That's an interesting proposal that I had not heard before.
12:10 p.m.
Assistant Professor, Department of Communication, University of Ottawa, As an Individual
One potential problem with requiring the registration and provision of a phone number is that a lot of mobile applications now connect to people's cellphone numbers, which means that, by allowing or forcing a phone number to be provided, there are potential ramifications for all kinds of different data joining, which would go beyond the ideal of.... A politician or candidate should be able to have that direct communication with constituents and try to engage them in the electoral process, which I think is really good, but I worry about creating a situation where people unknowingly give candidates access to all kinds of information from dating apps to Facebook to their Airbnb listings.
It's not necessarily that the specific listings are going to show up in a candidate's hands, but I think the possibilities for data joining when you use the cellphone number needs investigation. I'm not saying it's necessarily a bad idea, but it's not one I had considered before. My immediate gut reaction is that we need to be worried about the way different datasets can be connected together.
I think your basic point that candidates need to be able to connect with citizens is an important point and, in particular, candidates who are not incumbents end up at a disadvantage if they don't have good enough data to do that basic contact.