Snap Knew of Sextortion Risks, State Lawsuit Alleges
WSJ Tech News BriefingOctober 03, 202400:13:29

Snap Knew of Sextortion Risks, State Lawsuit Alleges

A lawsuit brought by New Mexico claims Snap failed to disclose or effectively address risks of sextortion, in which predators solicit explicit photos and then use them to blackmail the sender. WSJ reporter Jeff Horwitz tells host Zoe Thomas about the lawsuit’s allegations. Plus, venture capital pioneer Vinod Khosla says AI will drive deflation, but the productivity gains will allow governments to create social safety nets. Sign up for the WSJ's free Technology newsletter. Learn more about your ad choices. Visit megaphone.fm/adchoices

A lawsuit brought by New Mexico claims Snap failed to disclose or effectively address risks of sextortion, in which predators solicit explicit photos and then use them to blackmail the sender. WSJ reporter Jeff Horwitz tells host Zoe Thomas about the lawsuit’s allegations. Plus, venture capital pioneer Vinod Khosla says AI will drive deflation, but the productivity gains will allow governments to create social safety nets.


Sign up for the WSJ's free Technology newsletter.

Learn more about your ad choices. Visit megaphone.fm/adchoices

[00:00:18] [SPEAKER_02]: Welcome to Tech News Briefing, it's Thursday, October 3.

[00:00:23] [SPEAKER_02]: I'm Zoe Thomas for The Wall Street Journal.

[00:00:26] [SPEAKER_02]: A Silicon Valley pioneer says artificial intelligence will drive down the costs of services, ushering

[00:00:33] [SPEAKER_02]: in an age of abundance that can support a robust social safety net, if we let it.

[00:00:39] [SPEAKER_02]: We'll find out about Vinod Khosla's take on AI's potential impact.

[00:00:45] [SPEAKER_02]: And then, how do you warn users about the risks of sending explicit photos without scaring

[00:00:51] [SPEAKER_02]: them off your platform?

[00:00:53] [SPEAKER_02]: That's an issue Snap grappled with and failed to effectively address, according to an unredacted

[00:01:00] [SPEAKER_02]: version of a lawsuit brought against the company by New Mexico.

[00:01:04] [SPEAKER_02]: The state says Snap put growth over safety and failed to disclose design features that

[00:01:09] [SPEAKER_02]: make its platform a haven for abuse and sextortion.

[00:01:13] [SPEAKER_02]: Our reporter Jeff Horwitz joins us to talk about the case.

[00:01:20] [SPEAKER_02]: First up, venture capital pioneer Vinod Khosla is all in on AI.

[00:01:26] [SPEAKER_02]: Here's Khosla on The Wall Street Journal Pro's Executive Insights podcast, available on WSJ

[00:01:32] [SPEAKER_02]: Special Access.

[00:01:33] [SPEAKER_03]: Vinod Khosla, Executive Insights Podcast, Executive Insights Podcast, Executive Insights

[00:01:33] [SPEAKER_03]: Most expertise in the world, whether you're talking about structural engineers or oncologists,

[00:01:40] [SPEAKER_03]: mental health therapists or primary care doctors or journalists, teachers, that expertise will

[00:01:48] [SPEAKER_03]: be near free.

[00:01:49] [SPEAKER_02]: That voice you hear interjecting is Stephen Rosenbusch, one of our WSJ Pro Bureau chiefs.

[00:01:55] [SPEAKER_02]: He spoke with Khosla and Stephen is with us now.

[00:01:58] [SPEAKER_02]: So tell us more about Vinod Khosla.

[00:02:01] [SPEAKER_02]: Put his perspective on AI into context for us.

[00:02:05] [SPEAKER_04]: Vinod Khosla has been a force in Silicon Valley almost since it came into being.

[00:02:11] [SPEAKER_04]: He was the co-founder of Sun Microsystems in the early 1980s and he was an investor

[00:02:18] [SPEAKER_04]: at venture capital pioneer Kleiner Perkins in mid-1980s where he returned something on

[00:02:24] [SPEAKER_04]: the order of $10 billion to the firm.

[00:02:28] [SPEAKER_04]: And he's made a career out of taking massive all-in bets on new technology, new ideas

[00:02:37] [SPEAKER_04]: that sound improbable at the start.

[00:02:39] [SPEAKER_04]: He's willing to make these bets on areas where there's a high degree of failure but

[00:02:47] [SPEAKER_04]: the rewards of success are really, really outsized.

[00:02:51] [SPEAKER_04]: So if one thing doesn't work out, there are many, many other opportunities for him

[00:02:56] [SPEAKER_04]: to drive a huge gain that compensates for the many things that don't work out.

[00:03:02] [SPEAKER_04]: In this particular area, he is very, very confident though.

[00:03:08] [SPEAKER_02]: What is his take then on the pace of AI innovation?

[00:03:11] [SPEAKER_04]: He thinks that it's going to cause massive disruption.

[00:03:14] [SPEAKER_04]: About 10 years out, he thinks that AI will replace about 25 percent of today's current

[00:03:22] [SPEAKER_04]: labor.

[00:03:23] [SPEAKER_04]: At that level, we'll need a much, much broader social safety net.

[00:03:28] [SPEAKER_04]: But he argues that if we let AI essentially do its thing, that the productivity gains

[00:03:34] [SPEAKER_04]: from allowing AI to play a much bigger role in business and healthcare, in medical science,

[00:03:43] [SPEAKER_04]: in education will be so profound that productivity growth will begin to really accelerate and

[00:03:51] [SPEAKER_04]: we'll be able to afford a much broader and a much deeper social safety net than what

[00:03:57] [SPEAKER_04]: we have now or what's possible now.

[00:04:00] [SPEAKER_02]: Can you say a little bit more about that?

[00:04:01] [SPEAKER_02]: How does Koslis say this will drive down costs?

[00:04:05] [SPEAKER_04]: In his view, AI will take on the characteristics of what we consider an expert in today's world

[00:04:13] [SPEAKER_04]: and will come to rely on AI to provide not just productivity gains, but also advice,

[00:04:24] [SPEAKER_04]: services, judgment in areas from science to education to business.

[00:04:33] [SPEAKER_04]: This will allow us to reduce the cost of elder care.

[00:04:38] [SPEAKER_04]: It will allow us to reduce the cost of education because we'll have AI-based agents that will

[00:04:47] [SPEAKER_04]: be as good or even better than human teachers in the classroom in the not so distant future.

[00:04:57] [SPEAKER_04]: And that because the cost of labor in all of these areas will be driven down to something

[00:05:02] [SPEAKER_04]: close to zero and productivity will go up, we'll have an enormous amount of productivity

[00:05:10] [SPEAKER_04]: gains that can be reinvested elsewhere.

[00:05:14] [SPEAKER_02]: Is there anything that Koslis thinks could throw off this vision for how AI could sort

[00:05:20] [SPEAKER_02]: of fundamentally change our economy?

[00:05:22] [SPEAKER_04]: He thinks that what we do with AI is a political decision and a social decision more than a

[00:05:29] [SPEAKER_04]: technological decision.

[00:05:31] [SPEAKER_02]: All right, that was Stephen Rosenbusch.

[00:05:33] [SPEAKER_02]: He's a bureau chief for WSJ Pro.

[00:05:36] [SPEAKER_02]: Coming up, New Mexico claims Snap failed to warn users about sextortion risks on its platform.

[00:05:43] [SPEAKER_02]: We'll tell you about the state's case and Snap's response to the lawsuit after the break.

[00:05:58] [SPEAKER_02]: Snapchat, a social media app known for its vanishing messages, is popular with teens.

[00:06:04] [SPEAKER_02]: But the app's parent company, Snap, grappled internally with how to warn of sextortion

[00:06:11] [SPEAKER_02]: risks on its platform without striking fear in users and their parents.

[00:06:16] [SPEAKER_02]: That's according to an unredacted version of a lawsuit New Mexico filed against the

[00:06:21] [SPEAKER_02]: company in September.

[00:06:22] [SPEAKER_02]: Here to tell us more is our reporter Jeff Horwitz.

[00:06:25] [SPEAKER_02]: So Jeff, tell us more about what Snap allegedly knew about the risks of sextortion and how

[00:06:31] [SPEAKER_02]: it supposedly knew it.

[00:06:33] [SPEAKER_00]: Snap is a mixture of direct messaging with pictures and augmented reality features and

[00:06:39] [SPEAKER_00]: then a little bit of the regular social media feed that you're used to from Facebook and

[00:06:44] [SPEAKER_00]: other things.

[00:06:44] [SPEAKER_00]: And so it turns out because of its emphasis on person-to-person communication and because

[00:06:51] [SPEAKER_00]: of its emphasis on ephemeral communications, meaning that messages that literally disappear

[00:06:57] [SPEAKER_00]: and because of its young user base, that it was a particularly attractive forum for sextortion.

[00:07:04] [SPEAKER_00]: Basically, this would be someone convincing a person, frequently a minor, to share nude

[00:07:11] [SPEAKER_00]: pictures and then blackmailing them.

[00:07:14] [SPEAKER_00]: And this is something that every social media platform deals with.

[00:07:18] [SPEAKER_00]: But Snap was discovering that they were getting 10,000 reports a month and they were trying

[00:07:24] [SPEAKER_00]: to build up their defenses and also figure out what was going wrong and why their platform

[00:07:30] [SPEAKER_00]: seemed to be more vulnerable to this than others.

[00:07:34] [SPEAKER_02]: Despite the company's concerns about sextortion, Snap told The Wall Street Journal that sexting

[00:07:39] [SPEAKER_02]: accounts for a minuscule fraction of the messages exchanged on the platform.

[00:07:44] [SPEAKER_02]: But Jeff, what did Snap's internal reports say about how well the company was handling

[00:07:49] [SPEAKER_02]: this risk?

[00:07:50] [SPEAKER_00]: In terms of moderation and just responding, not great.

[00:07:55] [SPEAKER_00]: And I've written about stuff similarly with Meta with this.

[00:07:58] [SPEAKER_00]: They were pretty much failing all the time in a sample of cases when sextortion was reported

[00:08:04] [SPEAKER_00]: to them through the official Snap channels.

[00:08:06] [SPEAKER_00]: This is not unusual.

[00:08:08] [SPEAKER_00]: But so setting that aside, they were also looking at their product and trying to figure

[00:08:13] [SPEAKER_00]: out what about its product features might be making Snap a bit more of a home to people

[00:08:19] [SPEAKER_00]: who were bent on committing abuse and also why kids felt comfortable taking pictures

[00:08:27] [SPEAKER_00]: and sending them via Snap that they might not be willing to send elsewhere.

[00:08:32] [SPEAKER_02]: What did Snap find according to the documents cited in New Mexico's lawsuit?

[00:08:36] [SPEAKER_00]: So a few things.

[00:08:37] [SPEAKER_00]: Per New Mexico, the disappearing messages actually may have lulled users into a false

[00:08:43] [SPEAKER_00]: sense of confidence that this wasn't going to stick around and bite them.

[00:08:47] [SPEAKER_00]: And Snap did create a feature that alerted users when someone took a screenshot of their

[00:08:52] [SPEAKER_00]: phone.

[00:08:53] [SPEAKER_00]: Thing is, that doesn't always work depending on how that screenshot is recorded.

[00:08:57] [SPEAKER_00]: And also it's only after the fact, right?

[00:08:59] [SPEAKER_00]: So getting notified after someone already has compromise on you is not good.

[00:09:04] [SPEAKER_00]: So New Mexico is alleging basically that Snap failed to disclose this, that the product

[00:09:08] [SPEAKER_00]: features were sort of high risk and that the company didn't adequately tell teenagers and

[00:09:14] [SPEAKER_00]: their parents about these things.

[00:09:15] [SPEAKER_02]: What did New Mexico find when their investigators went and made a fake profile of a 14 year

[00:09:22] [SPEAKER_02]: old?

[00:09:22] [SPEAKER_00]: Yeah, so they found that commitments that are sort of what Snap had said publicly in

[00:09:27] [SPEAKER_00]: terms of how things worked and what safety features they had in place, that there were

[00:09:32] [SPEAKER_00]: at the very least significant exceptions to that.

[00:09:35] [SPEAKER_00]: For example, Snap was saying that this quick ad feature where they basically recommend

[00:09:40] [SPEAKER_00]: people that might be in your orbit to you to connect with, that it was in fact recommending

[00:09:46] [SPEAKER_00]: strangers and in fact some definitely inappropriate strangers to kids and to vice versa.

[00:09:53] [SPEAKER_00]: So it wasn't working.

[00:09:54] [SPEAKER_00]: And in fact, per New Mexico's discovery, Snap realized this in 2023, that in fact what they'd

[00:10:00] [SPEAKER_00]: been saying about their precautions, it wasn't working the way it was supposed to.

[00:10:05] [SPEAKER_00]: The company says, by the way, that it has worked to fix that, to address that and continues

[00:10:09] [SPEAKER_00]: to work on this stuff.

[00:10:11] [SPEAKER_02]: What has Snap said in particular about this lawsuit?

[00:10:14] [SPEAKER_00]: Snap has noted that the platform does have certain restrictions that are very much in

[00:10:23] [SPEAKER_00]: keeping with attempting to prioritize safety over growth.

[00:10:26] [SPEAKER_00]: For example, unlike Instagram or a lot of Twitter, you can't look and see everyone that someone's

[00:10:32] [SPEAKER_00]: connected to.

[00:10:33] [SPEAKER_00]: There are ways that you can, particularly if a user isn't trying very hard not to be

[00:10:39] [SPEAKER_00]: found.

[00:10:40] [SPEAKER_00]: There are ways that you can find people on there, but it's harder.

[00:10:44] [SPEAKER_00]: And Snap does view itself as being unfortunately the recipient of some of the bad behavior

[00:10:50] [SPEAKER_00]: and perhaps worse design features that are on other platforms.

[00:10:55] [SPEAKER_00]: Basically that by creating a place where kids feel comfortable and confident to some degree

[00:10:59] [SPEAKER_00]: that they end up being a venue of choice when people who are malicious are trying to get

[00:11:05] [SPEAKER_00]: kids to switch over to a different platform.

[00:11:07] [SPEAKER_02]: And where do things stand with the case?

[00:11:10] [SPEAKER_00]: I mean, this just got unredacted.

[00:11:12] [SPEAKER_00]: The case was filed last month, but this is the first time we were able to see what New

[00:11:16] [SPEAKER_00]: Mexico actually had from internal documents.

[00:11:19] [SPEAKER_00]: And a big question is whether these cases are designed in a way that's going to get

[00:11:24] [SPEAKER_00]: them past the traditional liability protections that platforms have.

[00:11:28] [SPEAKER_00]: Obviously, New Mexico is coming at this in a way that is intended to bypass that roadblock.

[00:11:36] [SPEAKER_02]: All right, that was our reporter Jeff Horwitz.

[00:11:39] [SPEAKER_02]: And that's it for Tech News Briefing.

[00:11:42] [SPEAKER_02]: Today's show was produced by Julie Chang with supervising producer Kathryn Millsop.

[00:11:46] [SPEAKER_02]: I'm Zoe Thomas for The Wall Street Journal.

[00:11:49] [SPEAKER_02]: We'll be back this afternoon with TNB Tech Minute.

[00:11:52] [SPEAKER_02]: Thanks for listening.