Information & Ethics Committee on May 28th, 2019

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

Thank you.

[The Chair, Conservative, Bob Zimmer]

Thank you, Mr. Angus.

Now we'll go to our delegations.

We'll go to the Parliament of Singapore for five minutes.

[Edwin Tong, Senior Minister of State, Ministry of Law and Ministry of Health, Parliament of Singapore]

Thank you very much for having me.

Thank you very much to all of you for your presentations this morning.

Mr. McNamee, you made the point that the business model of these platforms is really focused on algorithms that drive content to people who think they want to see this content. You also mentioned that fear, outrage, hate speech and conspiracy theories sell more. I assume by that you mean they sell more than truth does. Would that be right?

[Roger McNamee]

There was a study done at MIT in Cambridge, Massachusetts, that suggested that disinformation spreads 70% further and six times faster than fact. There are actually good human explanations for why hate speech and conspiracy theories move so rapidly. It's about triggering the flight or fight reflex.

[Edwin Tong]

Yes. When you throw in what Ms. Ressa said earlier about how disinformation is spread through the use of bots—I think she said that 26 fake accounts translated into three million different accounts that spread the information—I think we are facing a situation where disinformation, if not properly checked, goes exponentially viral. People get to see it all the time, and over time, unchecked, this leads to a serious erosion of trust and a serious undermining of institutions so that we can't trust elections, and fundamentally, democracy becomes marginalized and eventually demolished.

Would that be right, in your assessment?

[Roger McNamee]

I agree with that statement completely. To me, the challenge is in how you manage it. If you think about it, censorship and moderation were never designed to handle things at the scale that these Internet platforms operate at. In my view, the better strategy is to do the interdiction upstream, to ask the fundamental questions of what role platforms like this have in society and what business model is associated with them. To me, what you really want to do....

My partner, Renée DiResta, is a researcher in this area. She talks about the issue of freedom of speech versus freedom of reach, the latter being the amplification mechanism. On these platforms, what's really going on is the fact that the algorithms find what people engage with and amplify that more. Sadly, hate speech, disinformation and conspiracy theories are, as I said, the catnip that really gets the algorithms humming and gets people to react. In that context, eliminating that amplification is essential.

But how will you go about doing that, and how will you essentially verify that it's been done? To my mind, the simplest way to do that is to prevent the data from getting in there in the first place.

[Edwin Tong]

The point is that I think you must go upstream to deal with it, fundamentally, in terms of infrastructure. I think some witnesses also mentioned that we need to look at education, which I totally agree with, but when it does happen, and when you have the proliferation of false information, there must be a downstream or an end result kind of reach.

That's where I think your example of Sri Lanka is very pertinent, because it demonstrates that, left unchecked, the platforms would do nothing about the false information that goes around. What we do need is to have regulators and governments clothed with powers and levers to intervene swiftly and to disrupt the viral spread of online falsehoods very quickly.

Would you agree?

[Roger McNamee]

As a generalization, I would not be in favour of the level of government intervention that I have recommended here. I simply don't see alternatives at the moment.

In order to do what Shoshana is talking about and in order to do what Jim is talking about, you have to have some leverage, and the only leverage governments have today is their ability to shut these things down. Nothing else works quickly enough.

[Edwin Tong]

Yes, exactly. Speed is crucial in that situation. Look at what has happened in Sri Lanka, Myanmar and recently in Jakarta. That's what has happened.

[Roger McNamee]

Yes.

[Edwin Tong]

Thank you. My colleague will have some questions.

[Sun Xueling, Senior Parliamentary Secretary, Ministry of Home Affairs and Ministry of National Development, Parliament of Singapore]

I have some follow-up questions for Mr. McNamee. I'd like to make reference to the Christchurch shooting on March 15, 2019. After that, The New York Times published an article by Mr. Kevin Roose. I'd like to quote what he mentioned in his article. He said:

...we do know that the design of internet platforms can create and reinforce extremist beliefs. Their recommendation algorithms often steer users toward edgier content, a loop that results in more time spent on the app, and more advertising revenue for the company.

Those were his words. Would Mr. McNamee agree that the design of Internet platforms makes it easier for extremist views to thrive and gain a following?

[Roger McNamee]

Not only do I agree with that, I would like to make a really important point, which is that the design of the Internet itself is part of the problem. I'm of the generation—as Jim is as well—that was around when the Internet was originally conceived and designed. The notion in those days was that people could be trusted with anonymity, and that was a mistake, because bad actors use anonymity to do bad things. The Internet has essentially enabled disaffected people to find each other in a way in which they could never find each other in the real world and to organize in ways they could not in the real world.

When we're looking at Christchurch, we have to recognize the first step, which is that this was a symphonic work. This man went in and organized at least 1,000 co-conspirators prior to the act, using the anonymous functions of the Internet to gather them and prepare for this act. It was then, and only then, after all that groundwork had been laid, that the amplification processes of the system went to work. Keep in mind that those same people kept reposting the film. It is still up there today.

[Sun Xueling]

Yes, indeed.

[The Chair, Conservative, Bob Zimmer]

Thank you. We're past time, so if we have some time to get back around to finish your question, we will.

Next, we'll go to the Republic of Germany for five minutes.

[Jens Zimmermann, Social Democratic Party, Parliament of the Federal Republic of Germany]

Thank you very much, Mr. Chair.

I would like to start with Jim Balsillie. You mentioned in one of your six recommendations the question of taxation. As a member of our finance committee, I will say that this is, in many areas, an important aspect. Can you go a little deeper?

You mentioned especially the question of taxation of advertising. Do you see more areas there? Especially in the digital world, we know the problem of the shifting between countries and how difficult it is for countries to do a proper taxation.

[Jim Balsillie]

Sure. I'm talking about those who are buying the ads. Really, the core problem here is that when they're ad driven—you've heard extremely expert testimony on this—they'll do whatever it takes to get more eyeballs. The subscription-based model is a much safer place to be, because it's not attention driven.

One of the purposes of a tax is to manage externalities. If you don't like the externalities we're grappling with and that are illuminated here, then disadvantage those. Many of these platforms are moving more towards subscription-based models anyway, so just use tax as a vehicle to do that. The good benefit is that it gives you revenue. The second thing it could do is also begin to shift toward more domestic services. I think a tax has not been a lever that's been used, and it's right there for you.

[Jens Zimmermann]

Thank you. Maybe I'll make one comment on accountability.

From the experience we had in Germany with the introduction of our law, the so-called NetzDG, I would say that the aspect of accountability is one of the frequently mentioned aspects from which the networks have a lot of headaches, because at that point it's really getting personal, so I completely agree with what you said.

Professor Zuboff, I would like to ask you about the support of new forms of citizen action, which you mentioned. Being elected officials here, we know exactly how important it is to also convince the public, our voters. Also, from our experience in Germany, we know that many users have a lot of fear of something like what Roger McNamee mentioned—that we will shut it down. I don't know if we would exactly earn a lot of praise for doing that. We would have a lot of problems with our citizens.

How would you say we could encourage the users—the citizens—that some steps like these are needed? How can we avoid having that perceived as some sort of censorship by governments? We are always in that area where government interference can also be perceived as censorship.

[Shoshana Zuboff]

It's so critical to be conscious of that balance. Obviously, authoritarian governments would love Roger's recommendation—shut it down, because we don't like what they're saying. Obviously, that's not the intention here, so how do we make that distinction?

One thing I can say is that I really think we are in the midst of a sea change in this public reaction. I wonder if you're seeing this in Germany. I've been travelling all over the world, to many cities, over the last five months, continuously. With every group I talk to, I begin with one question: What are the concerns that brought you here?

In all different parts of the world and in every single group, no matter where I am, they say the same things. I ask them to shout out one word. It begins with “anxiety”, “manipulation”, “control”, “fear”, “resistance”, “democracy”, “freedom”, “rebellion”, “malaise”—the same constellation. What I've learned is that there is a sense, within our populations, that things are not right, that there is a power that is not aligned with our interests, that we don't understand it and that no one can control it.

That is beginning. With Cambridge Analytica, with Chris Wiley, our work is all making a difference. I think there is a ripeness there.

My advice would be to look to those areas where these new crystallizations are already emerging. Barcelona is one, which is based entirely on citizen solidarity. There are other cities as well that are getting on that bandwagon. There are groups of digital workers who are trying to devise digital communities and digital sovereignty.

It's about amplifying these things that are already coming up from the grassroots. The other side—and Maria was mentioning this as well—is education. We're still in a situation where every piece of peer research shows us, over and over again, that so many people simply do not understand these backstage operations. Why? Because billions of dollars have gone into designing them to keep us ignorant.

We have to break that, and we have to communicate and educate.

[The Chair, Conservative, Bob Zimmer]

Thank you very much.

Next up, we have the Republic of Chile.

Apparently the representative isn't here yet, so we'll go next to the representative from Estonia, for five minutes.

[Keit Pentus-Rosimannus, Vice-Chairwoman, Reform Party, Parliament of the Republic of Estonia (Riigikogu)]

Thank you.

Thank you very much for those inspiring presentations to kick off the morning.

I will start with a question to Mr. McNamee. My reading and my understanding are that it is really difficult to force the toothpaste back into the tube once it is out. I do think that the use of artificial intelligence—algorithms—is here to stay. To be very fair, AI is not evil per se.

I would put my question this way: If you were sitting in my chair today, what would be the three steps you would recommend, or would take, if we leave shutting down the platforms aside for a second?

[Roger McNamee]

The issue that we're dealing with here is that in the United States, or in North America, roughly 70% of all the artificial intelligence professionals are working at Google, Facebook, Microsoft or Amazon. To a first approximation, they're all working on behavioural manipulation. There are at least a million great applications of artificial intelligence. Behavioural manipulation is not on them. I would argue that it's like creating time-release anthrax or cloning human babies. It's just a completely inappropriate and morally repugnant idea, yet that is what these people are doing.

To Mr. Zimmermann's point, I would simply observe that it is the threat of shutting them down and the willingness to do it for brief periods of time that creates the leverage to do what I really want to do, which is to eliminate the business model of behavioural manipulation and data surveillance. I don't think this is about putting the toothpaste back in the tube. This is about formulating toothpaste that doesn't poison people.

I believe this is directly analogous to what happened with the chemical industry in the fifties. The chemical industry used to pour its waste products—mercury, chromium and things like that—directly into fresh water. They left mine tailings on the sides of hills. Petrol stations would pour spent oil into sewers, and there were no consequences so the chemical industry grew like crazy and had incredibly high margins. It was the Internet platform industry of its era. Then one day society woke up and realized that those companies should be responsible for the externalities that they were creating. That is what I'm talking about here.

This is not about stopping progress. This is my world. This is what I do. I just think we should stop hurting people. We should stop killing people in Myanmar, in the Philippines, and we should stop destroying democracy everywhere else. We can do way better than that. It's all about the business model.

I don't want to pretend I have all the solutions. What we know is that the people in this room are part of the solution, and our job is to help you get there. Don't view anything I say as a fixed point. View this as something that we're going to work on together.

The three of us are happy to take bullets for all of you, because we recognize it's not easy to be a public servant with these issues out there. But do not forget you're not going to be asking your constituents to give up the stuff they love. The stuff they love existed before this business model. It will exist again after this business model.

[Keit Pentus-Rosimannus]

Coming from a country where basically all the life is all so digital, I very much agree that it doesn't mean stopping the progress.

I will continue now with Mr. Balsillie. You, several times, underlined the need to regulate the political micro-targeting, or the political parties' ads. I must bring the example from 2007 when Estonia was for the first time under a very massive and serious cyber-attack. The main target of this attack was not the government sector, but it was mainly the private sector. As I saw, a lot of damage can be done, targeting also anything but party politics.

Do you see that the regulations need to be different for political micro-targeting and all the other ads, or do you see that basically the rules are needed the same way in both sectors?

[Jim Balsillie]

I agree with Roger that it needs to be in both sectors, but if there's one that's uniquely pernicious it's the underpinnings of our democracy. I think there needs to be complete transparency of all activities between political parties and these platforms. I think political parties should be under privacy legislation. Believe it or not, in Canada our political parties are not governed by our privacy legislation. I think it requires a special kind of personalized ban during elections.

I do agree with Roger that the business model is fundamentally flawed. It's going to take a constellation of activities to get there, but I think the political place is the most sensitive and most targetable place to fix it right away.

[Keit Pentus-Rosimannus]

A lot of damage can be done by micro-targeting ads concerning the environmental damage, for example, or medicine, or several other sectors.

My last question would be—