Evidence of meeting #153 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 facebook.
A video is available from Parliament.
11:50 a.m.
Conservative
The Chair Conservative Bob Zimmer
Next we will go to the Republic of Estonia.
Go ahead, for five minutes.
11:50 a.m.
Keit Pentus-Rosimannus Vice-Chairwoman, Reform Party, Parliament of the Republic of Estonia (Riigikogu)
Thank you.
I come from Europe, from Estonia. It is true that, two days after the European election, one can say that you actually have made progress in removing the fake accounts, but it is also true that those fake accounts should not have been there in the first place.
My first question is the following: What kinds of changes are you planning to use to identify your users in the beginning?
11:50 a.m.
Global Policy Director, Facebook Inc.
Thank you for that. I think we are cautiously pleased with the results in terms of the way the platform has performed in Europe. However, with respect to fake accounts, I would say—and I think my colleague Neil has mentioned—that this is a bit of an arms race with our adversaries, with bad actors trying to put inauthentic accounts and content onto the platform.
I think we want to constantly get better and to constantly evolve—
11:50 a.m.
Vice-Chairwoman, Reform Party, Parliament of the Republic of Estonia (Riigikogu)
How do you improve the identification process in the first place?
11:50 a.m.
Global Policy Director, Facebook Inc.
One of the things I can share with you, again in terms of the political advertising piece that we are trying to do, is that we are trying to get a very good certainty as to the identity of the individuals when they run an ad.
In Canada, as I mentioned earlier, we will do this and it will be not without pain. For political advertisers in Canada, it's very much that we will need to get some kind of ID from you. We will need to independently verify that ID, and then we're going to send you some kind of key—a digital key if you will—that you're going to have to use to authenticate yourself before you can even run a political ad.
This is a very costly and significant investment. It is also not without friction. I personally worry that there will be instances where people want to run an ad, don't know about this requirement and then find it's going to take many days to do.
I also worry that there may be false positives, but I think this is the right thing to do to protect elections around the world and here in Canada in October. We're prepared to put in the time, the investment and potentially some of the friction to get it right.
11:50 a.m.
Vice-Chairwoman, Reform Party, Parliament of the Republic of Estonia (Riigikogu)
All right. Thank you.
My second question will be about fake news or fake videos. What is your policy towards fake news, for example, deepfake? Will they be removed or will they just be marked as deepfake videos?
11:50 a.m.
Global Director, Information Policy, Google LLC
The issue of deepfakes.... Thanks for that question. It's a really important emerging issue. We have clear guidelines today about what content should be removed. If a deepfake were to fall under those guidelines, we would certainly remove it.
We also understand this needs further research. We've been working actively with civil society and academics on that.
11:50 a.m.
Vice-Chairwoman, Reform Party, Parliament of the Republic of Estonia (Riigikogu)
What about Facebook?
11:50 a.m.
Global Policy Director, Facebook Inc.
Thank you.
We are also investigating and doing research on this policy, to make sure we are in the right place. Currently, we would identify it as being fake and then inform our users, but we constantly—
11:50 a.m.
Vice-Chairwoman, Reform Party, Parliament of the Republic of Estonia (Riigikogu)
The deepfake will stay on.
11:50 a.m.
Global Policy Director, Facebook Inc.
We constantly iterate our policies, and we may update those policies in the future, as they evolve. We are working with research agencies and people on the ground to understand how these videos could manifest. As I mentioned before, if these videos, or any type of misinformation, led to real world harm—or off-line harm, I should say—we would remove that content.
11:50 a.m.
Vice-Chairwoman, Reform Party, Parliament of the Republic of Estonia (Riigikogu)
What about Twitter?
11:50 a.m.
Director, Public Policy, Twitter Inc.
We also share concerns about deepfakes. If we see the use of deepfakes to spread misinformation in a way that violates our rules, we'll take down that content.
11:50 a.m.
Vice-Chairwoman, Reform Party, Parliament of the Republic of Estonia (Riigikogu)
My last question will come back to what Mr. Collins asked in the beginning, about the fake video of Nancy Pelosi. Let's say a similar video were to appear, only the person there was Mr. Zuckerberg. Would that video be taken down, or just marked as a fake one?
11:50 a.m.
Voices
Oh, oh!
11:50 a.m.
Global Policy Director, Facebook Inc.
I'm sorry, the laughter.... I didn't hear the name?
11:50 a.m.
Vice-Chairwoman, Reform Party, Parliament of the Republic of Estonia (Riigikogu)
Sorry for the loud laughter. If a video similar to what has been airing, picturing Nancy Pelosi, appeared with Mark Zuckerberg, would you remove that fake video, or would you just mark it as fake news?
11:50 a.m.
Global Policy Director, Facebook Inc.
If it was the same video, inserting Mr. Zuckerberg for Speaker Pelosi, it would get the same treatment.
11:50 a.m.
Conservative
The Chair Conservative Bob Zimmer
Thank you.
I want to give a brief explanation of what's going on behind me. Your chairs have elected to have a working lunch. Feel free to come up and grab something to eat, and we'll continue with testimony all the way through our lunch.
Next up, we have Mexico. Go ahead for five minutes.
11:55 a.m.
Antares Guadalupe Vázquez Alatorre Senator
[Delegate spoke in Spanish, interpreted as follows:]
Thank you. I'm going to speak in Spanish.
I have several questions. In the case of Google, what do you do to protect people's privacy? I know many cases of sexting videos that are still there in the world, and when the victims of those videos go specifically to your office in Mexico—and I have known about several cases—the Google office tells them to go to the United States to complain. These are videos that are violently attacking somebody and cannot be downloaded. What do you do in those cases?
11:55 a.m.
Global Director, Information Policy, Google LLC
I'm not familiar with the particular cases you're talking about, but we do have strict guidelines about things like incitement to violence, or invasion of privacy and the like. If notified, and if we become aware of it, we would take action, if it's in violation of those guidelines. I'm not familiar with the specific cases, but would be happy to inquire further.
11:55 a.m.
Senator
Antares Guadalupe Vázquez Alatorre
[Delegate spoke in Spanish, interpreted as follows:]
I have information about very specific cases, but we don't know who to turn to in Mexico because they reject people. We hope you will take over this case, because the people have to wait for years to withdraw those things.
I'd like to ask Twitter about the creation of trends with robots. It's very common in Mexico. Every day, we have trends created with robots—the so-called bot farms. I don't know what the policy is at Twitter, because you seem to allow the artificial trends, or hashtags, when they are harming somebody. Why don't you allow trends to happen organically? I agree with having trends, things becoming viral and respecting freedom of expression, but why are robots allowed? If anybody can detect them, why is it Twitter cannot?
11:55 a.m.
Director, Public Policy, Twitter Inc.
Trends measure the conversation in real time and try to distinguish conversations that are always having a high level of engagement, English Premier League or the Mexican election writ large. What trends are trying to identify is acceleration above the normal. When that happens organically, like you mentioned, it has a different pattern from when it happens augmented by bots.
Since 2014, which for us is a lifetime, we've had the ability to protect trends from that kind of inorganic automated activity. Kevin mentioned an arms race. I think that's a good term for the battle against malicious automation. Right now we're challenging 450 million accounts a year for being inauthentic, and our tools are very subtle, very creative. They look at signals like instant retweets or activity that's so fast that it's impossible to be human.
Despite that, 75% is what ultimately gets kicked off the service, so 25% of the people who we thought were acting in an inauthentic way were able to pass the challenge. We are over-indexing to try to stop this inauthentic activity. This is a place where there is no delta between the societal values of trusting online activity and our imperatives as a company, which is that we want people when they come to Twitter to believe in what they see, to know that they are not getting messed about with Russian bots or whatever, so we work very hard to get this right and we're continuing to make improvements on a weekly basis.
11:55 a.m.
Senator
Antares Guadalupe Vázquez Alatorre
[Delegate spoke in Spanish, interpreted as follows:] But this happens normally every day, not during elections. It's happening every time that a trend is inflated with bots. There's been no adequate response to this. There are things that are aggressive.
This is for Twitter and for Facebook. When a user reports something that has been published, very commonly they report that they are not breaching their policy but they are aggressive against the person. They tell us that it's not breaching the policy, but it's somebody who is lying, who is attacking, and the person feels vulnerable. Many a time nothing happens because it's not breaching your policies.
Also, with this thing about the authentication, when the accounts are authenticated with blue arrows, even the other accounts can be blocked. Some people say identify yourself and they block all the accounts. Meanwhile, there are thousands of fake accounts and nothing happens, even if they are reported. There are users who are constantly reporting these fake accounts. Why do you have a different policy?