Information & Ethics Committee on Oct. 2nd, 2018

[Samantha Bradshaw]

I guess I'll just jump in here for the last point.

I definitely agree with Professor Dubois about the dangers of being able to link a phone number, say, to other datasets, because most people do connect their mobile to their Facebook profiles and other social media platforms. There is a real danger there that then you're starting to get more than just the phone number and basic contact information.

I think this does get to the root of one of the big problems we want to address here, and that is the kinds of data political parties should be allowed to use in the first place when they're campaigning and reaching out to voters. Then, what are the limits to that data? I think once we can come up with good answers to those questions, enforce them and create more transparency around this data that political parties have between them and the voters, then that will be a win for democracy.

[The Chair, Conservative, Bob Zimmer]

Thank you, Mr. Gourde.

Next up is Madam Fortier.

[Mona Fortier, Liberal, Ottawa—Vanier, ON]

Thank you very much.

I thank all three of you for being here today. We are working very hard on this study, and you have provided us with very relevant information for our work.

Mr. Pal, you mentioned that there were measures in Bill C-76 to prevent foreign or national interference in the elections. What effect could those changes have on third parties? That is one of my concerns.

During the recent election in Ontario, we saw Ottawa Proud, for instance, promote the information of a particular party, and not that of other parties. Do you think that an organization like that one could be affected by a foreign third party or by a government that has a lot of data in its possession?

How can we limit sharing large amounts of data with countries that have fewer regulations, even if financial exchanges are not necessarily involved?

Mr. Pal, I will let you begin, and the other two witnesses may speak afterwards.

[Prof. Michael Pal]

Thank you for your question. I'm going to answer it in English in order to be very precise.

Bill C-76 closes one of the loopholes that was still existing for foreign interference. It was already illegal for foreign entities to interfere in a Canadian election, but there was a loophole if you spent under $500. Five hundred dollars actually can get you a lot of Facebook ads in some markets. It would close that loophole. That's a very important measure.

How much does it cost to send mass WhatsApp messages or to advertise on Xbox? I don't know the answer to that. Those are emerging uses. Perhaps my colleagues have a sense on numbers. It's not clear how that affects other platforms, but certainly in terms of Facebook or Twitter advertising I think that helps.

I think the information is that third parties are playing an increasing role. There was a much larger number of third parties in the last federal election than there had been in previous elections. I think it's reasonable to expect that to continue. Third parties are basically only heavily regulated in terms of their spending during the election campaign, which is a relatively short period. Obviously you can try to influence voters before the official campaign starts, something that political parties have obviously realized, but also third parties.

I have been an advocate in my academic work for implementing pre-writ...so before the official election campaign rules, some of which are in Bill C-76. The spending limits would apply to what's called “partisan advertising” in the pre-writ period. I really welcome that. I actually think the pre-writ period should be longer. I said that to your colleagues at the procedure committee.

But we have a permanent campaign. Third parties have figured that out. There are other jurisdictions that take an even more aggressive.... Should there be disclosure of advertising by third parties in the full year leading up to the fixed election date? Should there be spending limits? I think there's a really good argument that there should be.

[Mona Fortier, Liberal, Ottawa—Vanier, ON]

I would like the other two witnesses to have time to answer. I thank you for having told us that.

In your presentation, you spoke about a study you submitted to MIT. Could you send that to the committee?

[Prof. Michael Pal]

That would be my pleasure, madam.

[Mona Fortier, Liberal, Ottawa—Vanier, ON]

Thank you.

Ms. Dubois, would you like to answer my question?

[Dr. Elizabeth Dubois]

Yes. I also will answer in English,

just so I can be precise.

I think that one of the major things is in addition to what Professor Pal has just put forward about the permanent campaign and where third parties are actually regulated. In addition to advertising and spending, we have questions about data collection itself. It's not just using the data to go and do the targeting, but it's collecting the data in the first place. We see a lot of this happening outside of the election period.

You mentioned Ontario Proud. They registered as an established third party but they weren't an established third party for a large chunk of their lifespan when they were collecting users on Facebook and updating their mailing lists and their text messaging list. That's all reasonable for a private entity or a non-profit organization, or whatever, to want to build up a contact list. But when they're doing that and then using it for very clear and explicit political purposes, that starts to raise questions about our ability to actually know whether or not the people who are in their database and then being sent political content and advertisements want to be there, whether or not the information is reliable, and whether or not the citizens would want to have their data removed. There's no real way that citizens are empowered to take ownership of their data.

[The Chair, Conservative, Bob Zimmer]

Thank you, Madam Fortier.

Next up we have Mr. Kent.

[Peter Kent, Conservative, Thornhill, ON]

Thank you, Chair.

When we talk about foreign intervention and potential foreign interference in the 2019 Canadian election campaign, we tend to think of Russia in our discussions. I'd like to ask some questions, however, about the history of the American-based Citizen Engagement Laboratory,which in 2015, in partnership with the Tides foundation, also based in the U.S., moved several million dollars into Canada to Leadnow, or through its Canadian subsidiary, the Tides subsidiary, the Sisu Institute Society in British Columbia.

You mentioned that $500 buys an awful lot of Facebook time, so does $1.7 million or $2 million. I'm wondering about the regulation. We're coming back to third party intervention, where the researcher in British Columbia, Vivian Krause, pointed out that it's easy for a political leader to take the high road in the campaign when a third party is well funded and is doing the dirty work on social media.

I wonder if all three of you could respond to that.

[Dr. Elizabeth Dubois]

I don't know the details of the Leadnow case. I know only what made the news headlines. I think, however, that it's a very important point when we're having these conversations about personal data use. With something like what C-76 has put forward for the requirements of a privacy statement, which already isn't very enforceable in the context of that bill, and which I think needs to change, it's very unclear how that then plays into the relationship between those parties and the third parties you've brought forward.

I think we need to be clearer about that. I don't have a specific recommendation on how to solve that problem, but I think it's a crucial one.

[Peter Kent, Conservative, Thornhill, ON]

This comes back to the question of foreign intervention.

[Prof. Michael Pal]

I don't have any information about that particular case. I believe C-76 requires third parties to have their own separate bank accounts, so that is one technical way of addressing the transfer of funds.

Constitutionally, you're on very solid ground to restrict spending. The Supreme Court has spoken about that. As for registration and transparency rules, the Supreme Court has spoken about that as well. There is also the question of contribution limits, in terms of donations to political parties and candidates. What's more controversial is whether you could or should have contribution limits for third parties.

Certainly B.C., in its referendum campaign on electoral reform, has limits on how much you can contribute to an advertising sponsor, and I believe this applies to their provincial elections as well. That does have an impact on political expression if you're restricting how much money.... Say a union or a corporation wants to spend its own money as a third party, how do you regulate that as a contribution? There are constitutional questions there but that's one way of addressing the movement of money between different entities, by treating it like a contribution.

[Peter Kent, Conservative, Thornhill, ON]

As we saw in the Brexit campaign, in the Leave campaign, it's possible to create subsets of the third party to get around those spending limits.

[Prof. Michael Pal]

If you have contributions being made in really large amounts and you have pre-restricted spending limits that are well enforced, then there's only so much money that can actually be used in that campaign. There's an interaction between the contribution limits and the spending limits.

[Samantha Bradshaw]

I'm not an expert on the legal points surrounding what you could actually do to help prevent foreign actors from making contributions to third parties or whatnot. What I do know and what I have seen in a lot of the research we've done tracking foreign influence in other countries is that they tend to work alongside a lot of the nationalist movements that are already in place.

Looking at the U.S., for example, it was always hard to differentiate the language of a Russian influencer from that used by members of alt-right organizations. A lot of that goes hand in hand and they work together to create somewhat of a shared narrative. I do think, however, that there are things we could do to regulate this problem. I think Professor Pal has made some good points. I don't think it's ever going to completely go away, but it's about raising the cost for these foreign and other bad actors. Raising the costs makes it just a little bit harder for them to start influencing voters.

I think transparency around where the funding is coming from, how much of it is spent, and what it's spent on would help create a little bit more accountability in the political system.

[The Chair, Conservative, Bob Zimmer]

Thank you.

Next up for five minutes is Ms. Vandenbeld.

[Anita Vandenbeld, Liberal, Ottawa West—Nepean, ON]

Thank you all of you for being here. Your testimony is incredibly enlightening.

I'd like to take a little step back and get my mind around what exactly we're talking about when we talk about bots, when we talk about cyborgs and when we talk about artificial intelligence. I think we all have a bit of an idea of what we mean but could we look at specific examples?

For example, Ms. Bradshaw, you said that you interviewed some of these bot developers. What does that mean—developing a bot? What goes into that? You said something about how they would have principles, like either virality or relevance, and that you could put principles that support democracy into a bot when you're developing it. Can you explain what exactly that means?

[Samantha Bradshaw]

Yes. When I was talking about the principles supporting the algorithms, I was talking more about social media algorithms and the way those things tailor content or deliver content to users. They deliver content based on virality. If things get a lot of clicks, that means that a lot of people find this really interesting and, therefore, it might trend. That's sort of what I was talking about there.

Instead of principles around virality, maybe we want principles on factual information coming from professional news outlets as opposed to sources that constantly produce misleading or fake news. The professional news should maybe get—I can't think of the technical word right now—but it should be prioritized in the algorithms. That's what I was referring to there when I was talking about designing algorithms for democracy. It's changing those sorts of principles.

When it comes to the actual bot developers.... This wasn't my core research project but I could point you to one of the researchers who did a lot more of these interviews than I did. His name is Sam Woolley. What he found was that the bot developers are just like any other tech developer. They're creating a piece of software that is designed to mimic human behaviour. It might amplify a certain story. It might converse with actual users online.

Bots do a whole bunch of different things. It really depends on what the goals of the developer are. The developers might actually have ideals or principles that they feed into the bots. A lot of them see these bots as being good for democracy because they're helping to amplify a message that might not get heard or go trending without the help of the bot.

[Anita Vandenbeld, Liberal, Ottawa West—Nepean, ON]

Ms. Dubois would like to speak.

[Dr. Elizabeth Dubois]

Political bots are automated accounts, sometimes found on social media, sometimes found on Instagram, sometimes found on Twitter, sometimes found in instant messaging apps. There's a bunch of different places where political bots are interacting.

There are other kinds of bots that make use of digital data to do things like make sure Wikipedia is up to date. There are different kinds of bots, but typically when I'm talking about political bots I mean these ones that are interacting with humans. They are mimicking humans in some way. The people creating those bots may have a lot of technical skill. They may be computer scientists and developers. They could also be people using tools that have been developed by some other developer to quickly create bots. You don't necessarily need to have a lot of technical skill to get a bot up and running.

We also need to remember that there are entire companies—in fact, an industry—built around the idea of search engine optimization that uses some of these techniques and could make use of a bot network they created. For example, there could be a whole bunch of bots they create say on Twitter to all interact with each other to amplify a message. There would be an organization potentially behind the development of those bots, and you can trickle down, as my colleague was explaining, to the specific person who was the original writer of the code.

[Anita Vandenbeld, Liberal, Ottawa West—Nepean, ON]

One of the things you mentioned, Ms. Dubois, was people who would game the algorithms. I think you even referred to breaking an algorithm. What does that mean?

[Dr. Elizabeth Dubois]

Gaming an algorithm would be.... The typical example is search engine optimization. Another way is, say, we have an event happening in Ottawa. We want the hashtag for that event to trend in Ottawa so that anybody in the Ottawa area sees the hashtag on their little trending bar on Twitter. The idea might be to create a whole bunch of bots that share content using that hashtag to artificially bump up how important it is on Twitter, so that Twitter's trending algorithm forces it to be front and centre for people to see. That would be a form of gaming the algorithm.

Breaking it would be gaming it to the point where the company that created the algorithm in the first place has to revamp it entirely because it's no longer doing what it's supposed to do.

[The Chair, Conservative, Bob Zimmer]

Thank you.

Ms. Vandenbeld, that's time.

We're going to do another round after this, but the last three-minute round goes to Mr. Angus.

[Charlie Angus, NDP, Timmins—James Bay, ON]

Thank you.

What started this whole investigation was the massive Facebook breach that led to Cambridge Analytica and the potential that that information undermined the Brexit vote. There was another Facebook breach of 50 million users. We have no idea. We're told not to worry. As far as they can tell, everything's fine.

As soon as I heard that, I thought, “Wow, thank God we have Facebook on the case. There's nothing to worry about here.”

When we had Facebook here, we were asking about the mass murders that happened in Myanmar. It's not the responsibility of Facebook that there were mass murders, but Facebook was accused time and time again of not responding to the misuse of their platform. Their response was something like, “We admit it, we're not perfect.” We're talking about the power of a platform to engage in mass killing.

We're talking about a lot of tweaks to a system that suddenly seems more powerful, more encompassing than domestic law, than anything we've dealt with in the past, and that seems to be moving beyond many jurisdictions with very little regard. Do you believe that platforms like Facebook, like Google, need to be regulated, or can we trust them to respond when there's enough outrage? Does there need to be antitrust action taken to break them up, since Facebook now controls Instagram, WhatsApp, and many other platforms? Google is the same.

What do you see in terms of holding these companies to account? Is it self-regulation? Is it antitrust? Is it some form of national or international regulation? I put that open.