After the Manchester attack, the UK is pushing Facebook to create a backdoor into WhatsApp encryption for terrorism investigations. Europe fines platforms up to €50 million for not removing extremist content within 24 hours, while Facebook claims to review 100 million pieces monthly. Keith Rabois argues that a technical backdoor, once created, cannot be reliably restricted and poses serious security risks.
Chapters
So, caroline, first of all, set the stage for us. How are the regulations in Europe different than they are in the United States?
Well, they've been talking a lot tougher. I mean, even in terms of bank for your back or bank for your euro, because in Germany they're even saying that we will find you social media outlets. Facebook included some €50 million. If within 24 hours you are not removing the hate speech or indeed the insightful videos that might be pointing towards extremist or or terrorist related activities, if you don't pull them down within 24 hours, you could be 50 million euros and indeed your executives could be fined €5 million a pop. So Germany has really led the charge ever since we look back to 2015 really when the immigration crisis when people from many a different country particularly Syria and Afghanistan came to Germany and this really incited a lot of hate speech at the time that actually saw people then congregating going out and struck violence to its heart. This got Germany active. France did the same. We heard from France and the representatives the social media giants that was back in 2015 after they had some deadly attacks. And now in the UK we're hearing the same sort of calls to calling on Twitter on Facebook to say be quicker at taking down this sort of insightful hate related content we need you to in the wake of the Manchester terror attacks.
Now Facebook says it reviews 100 million pieces of content every month. Is the scope of this problem just too big for them to deal with.
I think to manually review every piece of content absolutely is too big. I mean, there's a billion users per day or whatever there is and, you know, X number of posts, and I think that's unreasonable. Can you use math and algorithms to prioritize a very narrow set that needs to be reviewed? I think that's a scalable, you know, approach, and that can work very well. And I think they understand it very well, how to combine math and machines, math and machines with people in an efficient way.
And I think they'll be they're pretty there actually are better than they get.
Credit for it, and they'll be increasingly better. So you're optimistic about the crackdown on hate speech around Fake news?
No.
I mean, so I think there's very different things. I think fake news is actually a fake problem. And then I think, what do you mean by that? I don't believe there's any evidence that suggests that the average American, let's say, knew less in the last election about politics, policies, political figures than at any time in American history. I think in anything, the evidence is the contrary. People are more informed than ever before. So I don't think, and secondarily, I think fake news is an excuse that a lot of people use who used to be gatekeepers. to filter views and opinions, and they want to insert themselves back into the process of having a role of filtering. And the Internet as a whole has eliminated the role of gatekeepers and third parties in terms of filtering, and that's not stoppable. So I think that the fake news is a politically charged, politically biased approach to constraining a set of opinions. I think hate speech is well defined in the United States, it's generally defined fairly narrowly, and in Europe much more broadly. I think in the United States it would be very difficult to get tech companies to comply with laws passed by Parliament or some, you know, regulatory regime, because we have a very broad protection of the First Amendment that doesn't apply in Europe. And so the definitions of hate speech and incitement to terrorism or violence are much broader and much more wide than would ever be possible here to prevent.
Caroline, fake news or misinformation or whatever you want to call it was an issue following the Manchester terror attack. Now the U.K. government is pushing Parliament to pass a law that would require Facebook to basically remove encryption on WhatsApp messages belonging to suspects in investigations. Tell us the latest on this.
Yeah, this is fascinating. And of course, a bit of deja vu to the San Bernardino attacks in the United States. The same call is here that the United Kingdom Conservative Party looks which looks like it could well win by a landslide in the general election come June the 8th. They're calling for potentially the encryption that you have on WhatsApp, which is owned by Facebook, to that be basically a backdoor created, as was called after the San Bernardino shooting in the US to allow the authorities to access the data of those they believe are related to terrorist incidents. Now, it's interesting how this is all being digested by the United Kingdom at the moment, because many feel after the atrocities that occurred on Monday night in Manchester that, yes, it would be a good idea if we could ensure that those who have lethal views and might indeed use them can be accessed their data. Now, Rohan Silva, he's a big name in the UK tech tech community. He helped found Tech City with the previous Prime Minister David Cameron. He was one of those saying, look, actually, maybe this isn't a bad idea because in the United Kingdom, what it would be is a judge asking in specific circumstances to access the data of certain individuals. He says, look, maybe that's not a bad idea, tech companies, maybe you need to step up to the mark here. We're not asking for a blanket backdoor to be allowed, not allowed encryption to be eroded for all of us. but when it's needed by a judge, that could be the case.
How should, Keith, the tech companies approach these issues? Should they remove encryption in cases like this?
Well, I think it's a little bit more complicated. So, first of all, it's always been the case for 300 years that if you go to a judge, you get a warrant to search your house. Someone's always been able to get a warrant to go through all of your papers, all your belongings, your computers, everything at home. So the idea of applying the same principles to your phone is pretty trivial and not shouldn't be that controversial. The issue in the United States that a lot of people in the tech community have sort of published and broadcast is that once you create a backdoor, it's not clear that the government can constrain it and keep it in the right people's hands and that the wrong people don't all of a sudden have access to everybody's data. And that's like a technical argument. Is it possible to have a secure way to unencrypt things that can only be used by a very small number of people? And history sort of suggests maybe not. And math sort of suggests maybe not. So there's a danger that people are worried about, but the concept of a judge ordering access to data should be uncontroversial.
