Information & Ethics Committee on Oct. 2nd, 2018

[Samantha Bradshaw]

I'll jump in here first. For me, it's a little bit of everything. We definitely need regulation. If the past year has taught us anything it's that we need to step in, and government has a really important role to play in regulating these platforms.

I think private self-regulation is also an important thing to address here. For your example with Myanmar and how Facebook caused a lot of fake news and misinformation to spread, leading to violence and death, I think that's a real, serious problem. Like you said, we have these companies that operate globally, but they don't have staff working in each country on the content moderation side to address a lot of the very local problems.

Having a content moderator sitting in California who doesn't know anything about the history and culture of Myanmar or Sri Lanka, or a lot of these other countries where there are ethnic tensions, making decisions about content is a really big problem. Yet Facebook has advertising staff in a lot of these countries, so I think stepping up on their content moderation and making it more global and inclusive is a good private, self-regulatory step that governments could also push onto the platforms.

[Dr. Elizabeth Dubois]

I think self-regulation is absolutely insufficient. What Facebook in particular has been doing in reaction to the pressure now being put on them is good. We need to continue seeing those things.

Professor Pal brought up the example of making a repository of some election advertisements available voluntarily. Yes, that's great, and I am happy that happens, but as Professor Pal also mentioned, they could take that away and we would have absolutely no recourse. Elections Canada would then be left vulnerable because they decided to rely on something that was not legally mandated. If that's taken away during an election, we have a huge risk to our democracy and our democratic system.

We also need to remember that these are major international companies. They are not going to have the specifics of the Canadian population in mind when they're designing their self-regulation. Thinking about the Canadian population, we have large parts of the northern bit of Canada that are very reliant on Facebook as their main source of connection to political information. If Facebook is not actually able to pay attention to the nuances of aboriginal populations in Canada and the ways they share information, then those people are potentially underserved in a way that is counter to our democracy and potentially really marginalizing.

Why would we expect the people at Facebook who are making decisions about how Facebook will roll out all across the world to understand the specific Canadian context in those particular areas? That would be unreasonable, I think, to place on them and just say, “Oh yes, they'll take care of it.” In fact, we need the Canadian government to be the one standing up for Canadian citizens.

[The Chair, Conservative, Bob Zimmer]

We're at the end of our first round.

I'm going to ask a question to the panel, if I may.

One thing we've talked about in the past is anonymous accounts on Facebook and the platforms that.... They hide in the shadows. They're bots or they're real people who hide in the shadows, or whatever. Should it even be possible to hide in the shadows and not be a public entity? That to me seems like it would be an obvious place to go. An algorithm couldn't impersonate somebody who just isn't there, if they weren't there. It would have to be a real person attached to that particular file or particular program, etc., actually using it. Do you see that as even possible, or is that a place to go? What are your thoughts on that?

[Dr. Elizabeth Dubois]

One of the major problems with having what we would maybe call an “identity layer” like that is what becomes the verification of real personhood. Airbnb asks you to take a picture of your passport or your driver's licence. I'm not super comfortable with Airbnb having a copy of my passport when I logged in through Facebook and then 50 million user accounts' information has been accessed. There are major implications there that could be really problematic in terms of privacy and personal data security.

There are also questions about whether anonymity would actually fix some of these problems. Nathan—I forget his last name—he is is at MIT, wrote a really interesting review of the role of anonymity and whether it helps in cases of hate speech and disinformation. In fact, getting rid of anonymity largely wouldn't solve those problems, so that's potentially problematic as a solution.

[The Chair, Conservative, Bob Zimmer]

Ms. Bradshaw has something to add.

[Samantha Bradshaw]

I agree. I think there is a real risk in removing anonymity from these platforms. We do have to remember that they are global platforms. Protecting the identity of individuals in countries where there is less freedom of speech and less protection against violations of human rights.... People rely on anonymity on these platforms to communicate, to deliberate and to organize protests.

I could relate this to Twitter's little verified accounts. If, say, accounts were verified as being those of a real person, that could make fake accounts even more powerful, because they have now passed that filter and people will tend to trust those sources more. They could actually be more effective at spreading disinformation.

[The Chair, Conservative, Bob Zimmer]

That's interesting.

Mr. Pal.

[Prof. Michael Pal]

I think the proposition that anonymity can be important in facilitating political expression is a really good point. I think that applies in Canada too, in certain circumstances. I would draw the line where, if you're spending money on something that counts as election advertising under the definition of the Elections Act, then there should be some verification of the source, that it's a real person and a domestic actor who is behind it. It's different if you just want to express yourself, say, politically on Twitter and criticize a politician. There is value in allowing anonymity there.

That's where I would draw the line. That can be a difficult one to enforce, potentially. If you're spending money on advertising, then the public should have a right to know.

[The Chair, Conservative, Bob Zimmer]

Thank you.

First up for the next set of questions I have Mr. Baylis for seven minutes.

Go ahead, Mr. Baylis.

[Frank Baylis, Liberal, Pierrefonds—Dollard, QC]

I'm going to start exactly where you finished, Mr. Pal. I'd like all three of you to answer this. I'm going to start by trying to do what you said, Professor Dubois, define a political advertisement.

I see it two ways. If I've stayed in my house and written an article and put it out on Facebook or Twitter or whatever, and people start sharing it, whether I have 100 friends or 10,000 friends, I've just put it out there. If I take the same article and pay $10 to have it posted somewhere, I make the argument that this has now moved from a personal posting to an advertisement. I'd like to hear if this is a first line of delineation that we could do to say, “This is an ad. This is not an ad.”

Maybe, Professor Pal, you could start with that.

[Prof. Michael Pal]

There is a definition of advertising in the Elections Act: advocacy for or against a political party or a candidate, directly or indirectly, or on an issue they are associated with. We can work from there.

[Frank Baylis, Liberal, Pierrefonds—Dollard, QC]

But the delineation on being paid or not...?

[Prof. Michael Pal]

Some systems, a federal one, has a monetary amount, so Bill C-76 is going to eliminate that for foreign entities' advertising but it would still be there for domestic entities. B.C. did not have a monetary amount and there's a recent case in front of the Supreme Court of Canada trying to interpret what it means if you buy crayons and your kid puts up a political sign in the window, so having some kind of a monetary threshold I think is useful. It potentially stops over-enforcement of small political advertising.

[Dr. Elizabeth Dubois]

I think paid content versus unpaid content is a useful distinction. I think we also need to remember though that the content is important in deciding whether this is political or not. It's the dissemination of the content that makes it the advertisement. You could pay an artist to create a political campaign message that you leave in your house. That's paid and it's political but it's not being disseminated to others. The distinction there is important because things that help disseminate content that was not paid, I think, could reasonably still be considered an advertisement.

Imagine somebody writes a post on Twitter, they themselves don't pay to have it promoted, but somebody else chooses to promote it—

[Frank Baylis, Liberal, Pierrefonds—Dollard, QC]

I want to come back to that, but let's just start with them first of all.

The first act is that I write a post and I put it on there. I have not advertised if I've not paid.

[Dr. Elizabeth Dubois]

Right.

[Frank Baylis, Liberal, Pierrefonds—Dollard, QC]

I'm going to come back to that second point.

Professor Bradshaw.

[Samantha Bradshaw]

I won't add much to the conversation here other than that I think the differentiation between something that is paid for and something that I post organically is a good way to define an advertisement.

[Frank Baylis, Liberal, Pierrefonds—Dollard, QC]

Okay, then swing back to the second step. First of all, we say there's a monetary threshold. Now that threshold maybe should be brought down because its impact could be higher. We say, “This is a paid ad. This is not a paid ad.” Suddenly someone comes along with a bot and says this guy wrote a great article in his house and he shared it with 10 friends and we want to get this message out there. So the bot goes to work or I start paying. Then I've converted it into an ad at my stage, or even if I just told a bot to search the Internet, find this good stuff and then stick it out there, then that translates it from being a non-paid ad, when I put it up there, to now being computer driven.

Would that also be where we could say this is now a paid ad?

[Dr. Elizabeth Dubois]

Perhaps Professor Pal would have a better sense of whether under current law that would be the case. I think it should be, the caveat being it wouldn't be the person who created the content in the first place who should be held responsible, but instead, whoever has paid for those bots to be directed at that content.

[Frank Baylis, Liberal, Pierrefonds—Dollard, QC]

You're saying the person who has converted something into an ad.

[Dr. Elizabeth Dubois]

Yes.

[Frank Baylis, Liberal, Pierrefonds—Dollard, QC]

All right.

Professor Pal.

[Prof. Michael Pal]

Elections Canada has an interpretation note on political advertising online. It says that to count as election advertising, the item must have a placement cost. The legal question is whether there was money and transmission, but there also has to be a placement cost, which is different from, say, on television or radio. I guess the question is whether it counts as a placement cost or not.

We want to be cautious that we don't restrict the ability to share information organically.

[Frank Baylis, Liberal, Pierrefonds—Dollard, QC]

We're not sharing it organically. We're saying it has moved from being organic to being professionalized. Once a bot comes in or another person is paying, we've now moved it from organic to professionalized. In that instance, you're saying there is a placement cost.

Let's say I designed an app and I put it out there, and it was just doing its own thing without me. Would it be covered by this placement cost, or should we adapt the law?