WEBVTT - On with Kara Swisher: Deepfake abuse panel, featuring Laurie Segall

0:00:01.280 --> 0:00:04.600
<v Speaker 1>Hi, Mostly Human listeners, It's Lori Siegel. Today, I have

0:00:04.760 --> 0:00:05.640
<v Speaker 1>a treat for you.

0:00:06.040 --> 0:00:09.200
<v Speaker 2>I'm sharing an episode of Kara Swisher's podcast On with

0:00:09.280 --> 0:00:12.080
<v Speaker 2>Kara Swisher. In it, Kara hosts a panel of deep

0:00:12.119 --> 0:00:16.120
<v Speaker 2>fake experts, including me, to discuss the threat, the current scope,

0:00:16.160 --> 0:00:19.440
<v Speaker 2>and the future of sexually explicit defake abuse. Now, if

0:00:19.440 --> 0:00:21.680
<v Speaker 2>you've been listening to our four part series Searching for

0:00:21.720 --> 0:00:24.880
<v Speaker 2>Mister deep Fakes, this is a perfect cap. If you're

0:00:24.920 --> 0:00:26.880
<v Speaker 2>new to the subject, it's a fantastic primmer.

0:00:27.160 --> 0:00:30.479
<v Speaker 1>Hope you enjoy.

0:00:32.080 --> 0:00:35.559
<v Speaker 3>Hey, folks, award of warning before we begin. Today's episode

0:00:35.560 --> 0:00:40.360
<v Speaker 3>discusses sexual violence, including rape, graphic sexual content, and suicide,

0:00:40.520 --> 0:00:42.560
<v Speaker 3>so you might want to keep this in mind when

0:00:42.600 --> 0:00:57.760
<v Speaker 3>you decide if and where to listen. Hi, everyone from

0:00:57.800 --> 0:01:00.279
<v Speaker 3>New York Magazine and the Vox Media podcast Now Work.

0:01:00.520 --> 0:01:03.280
<v Speaker 3>This is on with Kara Swisher and I'm Kara Swisher.

0:01:03.600 --> 0:01:06.720
<v Speaker 3>My guests today are attorney Kerry Goldberg, tech journalists and

0:01:06.800 --> 0:01:11.080
<v Speaker 3>CEO of Mostly Human Media, Lori Siegel, and VS. Supermanian,

0:01:11.360 --> 0:01:15.759
<v Speaker 3>the Walter P. Murphy, Professor of Computer Science at Northwestern University.

0:01:16.160 --> 0:01:19.959
<v Speaker 3>We're talking about non consensual deep fake pornography or AI

0:01:20.040 --> 0:01:23.760
<v Speaker 3>generated explicit videos and images that use real people's likenesses

0:01:24.040 --> 0:01:27.400
<v Speaker 3>without their consent. Celebrities were often the target of early

0:01:27.480 --> 0:01:30.880
<v Speaker 3>pornographic deep fakes. Now, according to one survey, one in

0:01:30.959 --> 0:01:35.200
<v Speaker 3>eighteen's personally knows someone who has been targeted. Nearly all

0:01:35.280 --> 0:01:39.640
<v Speaker 3>deep fake porn depicts women. Despite laws and moderation efforts,

0:01:39.640 --> 0:01:43.680
<v Speaker 3>pornographic deep fake websites and so called newdify apps continue

0:01:43.720 --> 0:01:47.200
<v Speaker 3>to multiply, and a recent analysis from Wired found that

0:01:47.240 --> 0:01:51.360
<v Speaker 3>Elon musks rock chatbot is still being used to generate

0:01:51.440 --> 0:01:55.560
<v Speaker 3>and host non consensual deep fakes. I am not surprised

0:01:55.640 --> 0:01:59.240
<v Speaker 3>in the least. Carrie Goldberg is a victim's rights attorney.

0:01:59.280 --> 0:02:03.320
<v Speaker 3>Her law firm specializes in fighting Internet abuse and revenge. Porn.

0:02:03.680 --> 0:02:07.559
<v Speaker 3>Journalist Lori Siegel recently released Searching for Mister Deep Fakes,

0:02:07.560 --> 0:02:12.200
<v Speaker 3>an investigative docuseries created for TikTok and partnership with Paris Hilton.

0:02:12.600 --> 0:02:16.200
<v Speaker 3>It follows an investigation to unmask the anonymous operator behind

0:02:16.280 --> 0:02:20.080
<v Speaker 3>one of the Internet's most notorious deep fake porn platforms.

0:02:20.639 --> 0:02:24.760
<v Speaker 3>VS Subramanian is a top computer scientist and deep fake expert.

0:02:25.040 --> 0:02:27.880
<v Speaker 3>He led the team that developed the Global online deep

0:02:27.919 --> 0:02:32.399
<v Speaker 3>fake Detection system a free platform to help journalists detect

0:02:32.480 --> 0:02:35.320
<v Speaker 3>deep fakes. I think it's critically important to talk about

0:02:35.320 --> 0:02:38.120
<v Speaker 3>this issue. When I started my career, pornography was at

0:02:38.120 --> 0:02:41.000
<v Speaker 3>this heart of the problems with the early Internet and

0:02:41.080 --> 0:02:43.640
<v Speaker 3>the uses of it and the abuse of it was

0:02:44.080 --> 0:02:46.880
<v Speaker 3>built into the system at the start. A lot of

0:02:46.880 --> 0:02:48.560
<v Speaker 3>people were aware of it then and have had a

0:02:48.680 --> 0:02:51.280
<v Speaker 3>very hard time dealing with it even as technology has

0:02:51.320 --> 0:02:54.000
<v Speaker 3>gotten better and better at being more and more abusive.

0:02:54.040 --> 0:02:56.560
<v Speaker 3>So it's critical that we keep talking about it, we

0:02:56.680 --> 0:03:00.880
<v Speaker 3>criminalize this, and we start to hold platforms, big platforms

0:03:01.080 --> 0:03:05.240
<v Speaker 3>responsible for the stuff that they distribute. Our expert question

0:03:05.280 --> 0:03:08.760
<v Speaker 3>today comes from Wall Street Journal tech reporter Georgia Wells.

0:03:09.160 --> 0:03:11.840
<v Speaker 3>This is an important conversation and one we will keep having,

0:03:11.960 --> 0:03:15.160
<v Speaker 3>so stick around and one more thing before we get started.

0:03:15.160 --> 0:03:18.000
<v Speaker 3>If you're going to be in Washington, DC on July sixteenth,

0:03:18.000 --> 0:03:21.560
<v Speaker 3>please join us live at the Johns Hopkins University Bloomberg

0:03:21.600 --> 0:03:24.200
<v Speaker 3>Center for a taping up on I'll be talking to

0:03:24.240 --> 0:03:28.320
<v Speaker 3>Gina Rimando, Commerce Secretary under President Biden and former Rhode

0:03:28.320 --> 0:03:32.080
<v Speaker 3>Island governor about what AI means for the American workforce,

0:03:32.400 --> 0:03:35.160
<v Speaker 3>and as a special bonus before that, conversation. I'll be

0:03:35.200 --> 0:03:38.520
<v Speaker 3>speaking with Johns Hopkins University President Ron Daniels and the

0:03:38.600 --> 0:03:42.800
<v Speaker 3>University of Notre Dame President Father Robert Dowd about how

0:03:42.960 --> 0:03:46.760
<v Speaker 3>universities are approaching AI and workforce issues. You can get

0:03:46.800 --> 0:03:50.000
<v Speaker 3>tickets and learn more at boxmedia dot com. Slash Kara

0:03:50.120 --> 0:04:04.000
<v Speaker 3>Swisher Live, Carrie, Laurie and VS. Thanks for coming on on.

0:04:04.320 --> 0:04:05.000
<v Speaker 3>Could you be here?

0:04:05.320 --> 0:04:06.200
<v Speaker 4>Thrilled to be here?

0:04:06.560 --> 0:04:10.280
<v Speaker 3>Hello, Hello, So. I talked to experts back in January

0:04:10.360 --> 0:04:13.960
<v Speaker 3>about the proliferation of sexualized non consensual AI deep fikes,

0:04:14.000 --> 0:04:17.280
<v Speaker 3>including images of children on Elon muss x that were

0:04:17.360 --> 0:04:21.200
<v Speaker 3>generated by its groc chatbot. Obviously, AI created deep fake

0:04:21.279 --> 0:04:24.599
<v Speaker 3>porn hasn't gone away on Groc or elsewhere, and the

0:04:24.640 --> 0:04:27.400
<v Speaker 3>technology continues to improve. And like each of you to

0:04:27.440 --> 0:04:30.800
<v Speaker 3>explain the current scope of the problem and how pervasive

0:04:30.839 --> 0:04:35.280
<v Speaker 3>it is and who it's affecting, and how technically advanced

0:04:35.279 --> 0:04:37.200
<v Speaker 3>it is. We're going to get more into detail, So

0:04:37.279 --> 0:04:40.760
<v Speaker 3>go high level here, Carrie, We'll start with you, than Laurie,

0:04:40.960 --> 0:04:41.880
<v Speaker 3>than VS.

0:04:42.200 --> 0:04:42.320
<v Speaker 2>So.

0:04:42.880 --> 0:04:48.040
<v Speaker 5>I'm a plaintiff's attorney that sues tech companies, and generally

0:04:48.120 --> 0:04:52.479
<v Speaker 5>I see problems at the very early stages. So you know,

0:04:52.560 --> 0:04:55.960
<v Speaker 5>back probably seven years ago, we started getting situations where

0:04:56.160 --> 0:05:00.760
<v Speaker 5>celebrities and really high level content creators were deep faked.

0:05:01.200 --> 0:05:04.919
<v Speaker 5>And then over the last just i'd say year or so,

0:05:05.520 --> 0:05:09.000
<v Speaker 5>it started trickling down to middle schoolers getting deep faked.

0:05:09.160 --> 0:05:12.840
<v Speaker 5>But all the apps involved were these pretty obscure things.

0:05:12.880 --> 0:05:17.719
<v Speaker 5>But then everything totally changed when Elon started sending around,

0:05:17.760 --> 0:05:20.599
<v Speaker 5>you know, pictures of toasters in bikinis, and it just

0:05:21.000 --> 0:05:25.880
<v Speaker 5>unleashed this never before seen sort of situation where everyone

0:05:25.960 --> 0:05:29.200
<v Speaker 5>was able to create deep fakes and then immediately publish

0:05:29.240 --> 0:05:31.640
<v Speaker 5>them on the X platform.

0:05:31.640 --> 0:05:33.320
<v Speaker 3>So that's where you saw it, Laurie.

0:05:33.640 --> 0:05:36.000
<v Speaker 2>Yeah, when we started looking at this years ago, it

0:05:36.080 --> 0:05:41.160
<v Speaker 2>was very much focused on celebrities, oftentimes women in power politicians.

0:05:41.200 --> 0:05:42.599
<v Speaker 1>It seems like the first line of defense.

0:05:42.600 --> 0:05:45.080
<v Speaker 2>And I remember thinking at the time, oh, no, you know,

0:05:45.160 --> 0:05:47.040
<v Speaker 2>you might not care about these people. We should care

0:05:47.040 --> 0:05:49.240
<v Speaker 2>about these people. But what this means is this is

0:05:49.279 --> 0:05:52.160
<v Speaker 2>going to happen in our high schools, and unfortunately that's

0:05:52.160 --> 0:05:55.599
<v Speaker 2>what we're seeing happening. So I would just say it's rampant,

0:05:55.760 --> 0:06:00.359
<v Speaker 2>and it's happening in our communities. It's happening to women,

0:06:00.560 --> 0:06:03.800
<v Speaker 2>to all sorts of folks, and unfortunately it's something that

0:06:03.839 --> 0:06:06.640
<v Speaker 2>teenagers are learning and it's becoming commonplace.

0:06:06.920 --> 0:06:11.240
<v Speaker 4>Go ahead, yes, Kara, I'm a technologist, and so I've

0:06:11.279 --> 0:06:15.400
<v Speaker 4>spent a lot more time thinking about the technology over

0:06:15.440 --> 0:06:19.400
<v Speaker 4>the last you know, five six years ago, the ability

0:06:19.440 --> 0:06:23.760
<v Speaker 4>to create deep fakes was in its infancy, but over

0:06:23.760 --> 0:06:26.760
<v Speaker 4>the last five years, and in particular over the last

0:06:26.839 --> 0:06:29.640
<v Speaker 4>year or two, it's really gotten to a stage where

0:06:29.680 --> 0:06:33.240
<v Speaker 4>even experts like me on the detection of deep fakes

0:06:33.240 --> 0:06:38.160
<v Speaker 4>have difficulty eyeballing something, listening to something, and telling if

0:06:38.160 --> 0:06:40.520
<v Speaker 4>it's real or fake. So what we're seeing is something

0:06:40.560 --> 0:06:43.160
<v Speaker 4>where you really need humans and AI to come together

0:06:43.680 --> 0:06:47.200
<v Speaker 4>to work as some kind of partnership to try and

0:06:47.240 --> 0:06:50.640
<v Speaker 4>figure this out. And you know, until we can get

0:06:50.680 --> 0:06:52.920
<v Speaker 4>that right, I don't see an end in sight.

0:06:53.360 --> 0:06:56.640
<v Speaker 3>And how has the technology improved in the last year

0:06:57.279 --> 0:06:58.120
<v Speaker 3>Just ease of use.

0:06:58.720 --> 0:07:02.000
<v Speaker 4>Ease of use has always been there. The improvement has

0:07:02.000 --> 0:07:05.360
<v Speaker 4>been in the quality of the generator deep fakes. So today,

0:07:05.400 --> 0:07:09.280
<v Speaker 4>for example, we're seeing things like face swapping, so I

0:07:09.279 --> 0:07:15.280
<v Speaker 4>could record something of myself saying something thoroughly objectionable and

0:07:15.360 --> 0:07:18.680
<v Speaker 4>then put, you know, replace my face with that of

0:07:18.680 --> 0:07:22.040
<v Speaker 4>a celebrity in the past that was difficult to do.

0:07:22.560 --> 0:07:24.840
<v Speaker 4>There were things around the edges of the face that

0:07:24.920 --> 0:07:28.320
<v Speaker 4>would be noticeable. Today, much of that has been worked

0:07:28.320 --> 0:07:28.800
<v Speaker 4>out and.

0:07:28.880 --> 0:07:31.400
<v Speaker 3>Ease of use for normal people to do so correct.

0:07:31.440 --> 0:07:32.960
<v Speaker 4>Indeed, absolutely yeah.

0:07:33.000 --> 0:07:36.320
<v Speaker 3>So mister d Fakes launched in twenty eighteen and was

0:07:36.360 --> 0:07:39.440
<v Speaker 3>once again the largest deep fake porn site online, getting

0:07:39.480 --> 0:07:43.000
<v Speaker 3>seventeen million visits a month at its peak. Laurie, talk

0:07:43.000 --> 0:07:46.240
<v Speaker 3>about your investigation into mister deep Fakes and how it began.

0:07:47.000 --> 0:07:49.040
<v Speaker 2>I began a bit with an obsession, to be quite

0:07:49.040 --> 0:07:52.680
<v Speaker 2>honest with you. I remember it was twenty twenty two,

0:07:52.840 --> 0:07:56.080
<v Speaker 2>chatch Ept had launched. Everybody was talking about how AI

0:07:56.280 --> 0:07:58.160
<v Speaker 2>was going to just do these incredible things. It's going

0:07:58.200 --> 0:08:01.320
<v Speaker 2>to cure cancer, it's going to to you know, give

0:08:01.400 --> 0:08:02.480
<v Speaker 2>us all this free time.

0:08:02.600 --> 0:08:04.520
<v Speaker 1>And of course, I mean, car, you know this better

0:08:04.560 --> 0:08:04.920
<v Speaker 1>than anyone.

0:08:04.960 --> 0:08:06.920
<v Speaker 2>We're all kind of looking at this and saying, okay,

0:08:07.200 --> 0:08:09.400
<v Speaker 2>there are a lot of unintended consequences, and this is

0:08:09.440 --> 0:08:13.320
<v Speaker 2>moving very quickly without the correct guardrails. And someone told

0:08:13.360 --> 0:08:16.640
<v Speaker 2>me about mister deep Fakes. I will never forget going

0:08:16.640 --> 0:08:21.520
<v Speaker 2>to this site and opening it up and seeing thousands,

0:08:21.840 --> 0:08:26.600
<v Speaker 2>thousands of highly sexually graphic images, deep fakes of women

0:08:26.600 --> 0:08:29.480
<v Speaker 2>who did not consent, and they were violent in nature.

0:08:30.000 --> 0:08:33.080
<v Speaker 2>It was a website where a creator could say, you know,

0:08:33.440 --> 0:08:35.600
<v Speaker 2>could go and make money, and someone could go and

0:08:35.640 --> 0:08:38.439
<v Speaker 2>say I want this woman doing this, and they could

0:08:38.480 --> 0:08:38.920
<v Speaker 2>pay for it.

0:08:38.960 --> 0:08:41.119
<v Speaker 1>So it was beyond just a website. It was a platform.

0:08:41.400 --> 0:08:43.200
<v Speaker 2>And then what was even more concerning as we were

0:08:43.240 --> 0:08:46.400
<v Speaker 2>digging into it was, you know, there was everything from

0:08:46.600 --> 0:08:50.040
<v Speaker 2>training data of women's faces to create hyper realistic deep

0:08:50.120 --> 0:08:53.320
<v Speaker 2>fake pornography, and then there were forums where people would

0:08:53.360 --> 0:08:56.160
<v Speaker 2>exchange notes on here's the latest deep fake tools, here's

0:08:56.200 --> 0:08:58.720
<v Speaker 2>how you can do this. You can pay a premium

0:08:58.760 --> 0:09:01.720
<v Speaker 2>and get a deep fake training manual. And so my

0:09:01.800 --> 0:09:04.000
<v Speaker 2>initial thought was this isn't going to just be for

0:09:04.040 --> 0:09:06.520
<v Speaker 2>these high profile women and high profile I would say,

0:09:06.600 --> 0:09:08.760
<v Speaker 2>let's take it with a grain of salt. And I

0:09:08.800 --> 0:09:11.120
<v Speaker 2>remember thinking at the time, well, it's not going to

0:09:11.160 --> 0:09:13.760
<v Speaker 2>stop here. You know, this is just the beginning of it.

0:09:13.800 --> 0:09:15.760
<v Speaker 2>And by the way, you're looking at these forums and

0:09:15.800 --> 0:09:18.160
<v Speaker 2>you're seeing people say, is it wrong that I want

0:09:18.160 --> 0:09:20.360
<v Speaker 2>to do this to my sister in law? And oh,

0:09:20.440 --> 0:09:23.319
<v Speaker 2>this is so easy to do, and so it's almost

0:09:23.400 --> 0:09:27.320
<v Speaker 2>creating entertainment around deep fake abuse. And so as I

0:09:27.360 --> 0:09:30.080
<v Speaker 2>was looking at this, I thought, well, surely there's something

0:09:30.120 --> 0:09:33.920
<v Speaker 2>these women can do to help themselves, and the reality

0:09:34.000 --> 0:09:36.560
<v Speaker 2>was absolutely not. There was nothing they could do because

0:09:36.880 --> 0:09:39.559
<v Speaker 2>the man behind it was anonymous, and he'd been anonymous

0:09:39.600 --> 0:09:42.679
<v Speaker 2>for six years. And even so, even if he wasn't anonymous,

0:09:42.679 --> 0:09:45.400
<v Speaker 2>the laws hadn't caught up. And even when I went

0:09:45.400 --> 0:09:47.079
<v Speaker 2>out and started talking about this, and by the way,

0:09:47.080 --> 0:09:48.480
<v Speaker 2>you never want to be the person talking about deep

0:09:48.480 --> 0:09:50.520
<v Speaker 2>fake porn at the dinner table, but turns out I was.

0:09:50.880 --> 0:09:53.560
<v Speaker 2>People were like, well, the images aren't real that and

0:09:53.640 --> 0:09:57.200
<v Speaker 2>so it doesn't matter, and I'm sure everybody on here

0:09:57.280 --> 0:09:59.200
<v Speaker 2>is kind of just shaking their head. So we also

0:09:59.240 --> 0:10:02.120
<v Speaker 2>had a cultural problem. We needed people to realize that

0:10:02.120 --> 0:10:03.920
<v Speaker 2>these are so hyper realistic.

0:10:04.480 --> 0:10:05.920
<v Speaker 1>Actually this causes.

0:10:05.640 --> 0:10:08.520
<v Speaker 2>Deep, deep psychological abuse, and it can be used to

0:10:08.600 --> 0:10:11.840
<v Speaker 2>try to digitally abuse a woman, I say, a woman anyone,

0:10:12.280 --> 0:10:14.280
<v Speaker 2>And so that was what was alarming, and so we

0:10:14.360 --> 0:10:16.920
<v Speaker 2>decided we had to go find him and raise awareness.

0:10:16.960 --> 0:10:18.600
<v Speaker 1>So that was the beginning of the investigation.

0:10:19.240 --> 0:10:22.199
<v Speaker 3>Let's hear from Joanne Chu. She's an artist and actor.

0:10:22.360 --> 0:10:25.079
<v Speaker 3>One of the women featured in Laurie's investigation. She's talking

0:10:25.080 --> 0:10:27.600
<v Speaker 3>about the moment when she realized her image had been

0:10:27.720 --> 0:10:29.800
<v Speaker 3>used in porn on mister deep Fakes.

0:10:30.679 --> 0:10:33.040
<v Speaker 6>I was just bored one evening and I'm just gonna

0:10:33.040 --> 0:10:35.199
<v Speaker 6>google Google my name and see what comes up. I

0:10:35.320 --> 0:10:37.240
<v Speaker 6>kept scrolling down, and then I said.

0:10:37.040 --> 0:10:38.280
<v Speaker 4>What what what? What?

0:10:38.280 --> 0:10:38.520
<v Speaker 6>What?

0:10:38.520 --> 0:10:39.040
<v Speaker 4>What is this?

0:10:41.320 --> 0:10:45.439
<v Speaker 6>It's like my face on very uh disturbing and graphic

0:10:45.720 --> 0:10:51.760
<v Speaker 6>graphic content. It's you know, Joanne cho or Joanne Cheo.

0:10:52.679 --> 0:10:55.560
<v Speaker 6>In the beginning, it was just a few listings mister

0:10:55.559 --> 0:10:58.920
<v Speaker 6>deep fakes came up social media girls e Porner, and

0:10:58.960 --> 0:11:02.400
<v Speaker 6>then when I typed it in against it multiplied. It's

0:11:02.559 --> 0:11:06.560
<v Speaker 6>very much akin to physical assault. I have like less

0:11:06.559 --> 0:11:09.880
<v Speaker 6>than three thousand followers, but it doesn't matter if you're

0:11:09.960 --> 0:11:12.480
<v Speaker 6>famous or you're not. Just looking at mister dee fakes.

0:11:12.520 --> 0:11:15.720
<v Speaker 6>It's like they have a whole alphabetized database. At one point,

0:11:16.720 --> 0:11:18.920
<v Speaker 6>when I typed in my name, my stuff wasn't coming

0:11:18.960 --> 0:11:20.760
<v Speaker 6>up first anymore. It was all of this content.

0:11:21.280 --> 0:11:24.240
<v Speaker 3>So she compares to being physically assaulted. Carrie, does this

0:11:24.320 --> 0:11:26.680
<v Speaker 3>reflect what you've heard from your clients, because it sounds

0:11:26.720 --> 0:11:29.560
<v Speaker 3>like perpetrators are telling themselves that it isn't real harm.

0:11:30.160 --> 0:11:32.959
<v Speaker 5>Yeah, I remember this when I first started my firm

0:11:33.040 --> 0:11:36.520
<v Speaker 5>back in twenty fourteen, and it was everyone's attitude about

0:11:36.559 --> 0:11:39.640
<v Speaker 5>revenge porn and how you know, this is just an

0:11:39.720 --> 0:11:44.000
<v Speaker 5>online problem. And back then there were all these revenge

0:11:44.040 --> 0:11:47.839
<v Speaker 5>porn websites and people were searchable through Google, which would

0:11:47.960 --> 0:11:51.920
<v Speaker 5>lead people to the search engine results and it would

0:11:51.960 --> 0:11:55.640
<v Speaker 5>just be pages and pages of nude images. Like all

0:11:55.640 --> 0:11:58.880
<v Speaker 5>the harm was really really underestimated back then, and it

0:11:58.920 --> 0:12:04.280
<v Speaker 5>took a huge, huge amount of advocacy to change the

0:12:04.320 --> 0:12:07.880
<v Speaker 5>public opinion about the shame around being you know, nude

0:12:07.920 --> 0:12:11.840
<v Speaker 5>online and we're just we're seeing something really really similar.

0:12:12.000 --> 0:12:14.800
<v Speaker 5>The difference though, is that back you know, ten years ago,

0:12:15.400 --> 0:12:19.280
<v Speaker 5>women were being blamed for having shared an intimate image

0:12:19.280 --> 0:12:22.400
<v Speaker 5>in the first place, and with deep fakes, I don't

0:12:22.400 --> 0:12:25.880
<v Speaker 5>see that same kind of victim blaming because here no

0:12:25.960 --> 0:12:29.800
<v Speaker 5>one consent it to the image in the first place.

0:12:30.160 --> 0:12:33.920
<v Speaker 5>But there is still a lot of I think marginalization

0:12:34.240 --> 0:12:36.320
<v Speaker 5>and minimizing of the harm.

0:12:36.440 --> 0:12:38.559
<v Speaker 3>That it's not physical assault, that it's.

0:12:39.760 --> 0:12:43.360
<v Speaker 5>Yeah, I feel it is conceptually different than a physical assault,

0:12:43.480 --> 0:12:47.400
<v Speaker 5>but that doesn't mean that it's not still really really

0:12:47.679 --> 0:12:52.440
<v Speaker 5>harmful and humiliating for victims, especially like I see so

0:12:52.559 --> 0:12:57.240
<v Speaker 5>many young people who are like victims, and it's other

0:12:57.400 --> 0:12:59.600
<v Speaker 5>young people's doing it to them, you know, like where

0:12:59.600 --> 0:13:03.160
<v Speaker 5>it's on peer in junior high school and high school settings,

0:13:03.160 --> 0:13:05.320
<v Speaker 5>and a lot of times the offenders just think that

0:13:05.360 --> 0:13:09.680
<v Speaker 5>they're doing a prank or if you know, and always

0:13:09.760 --> 0:13:12.440
<v Speaker 5>the defense is, well, I barely shared it with anybody,

0:13:12.480 --> 0:13:15.720
<v Speaker 5>I didn't post it online. But these things find a

0:13:15.760 --> 0:13:17.000
<v Speaker 5>way to get published.

0:13:17.360 --> 0:13:20.600
<v Speaker 3>So deepig technology first to merge around twenty fifteen. VS

0:13:20.720 --> 0:13:24.240
<v Speaker 3>talk about the online communities that exist for sharing information

0:13:24.480 --> 0:13:28.400
<v Speaker 3>making so called newdified tools accessible and how has it

0:13:28.440 --> 0:13:31.199
<v Speaker 3>evolved over the past ten years in terms of how

0:13:31.240 --> 0:13:32.079
<v Speaker 3>it's used.

0:13:32.440 --> 0:13:36.120
<v Speaker 4>So Kara, there are a number of neutification AI tools.

0:13:36.679 --> 0:13:40.640
<v Speaker 4>So these are tools that are of various varieties. One,

0:13:40.760 --> 0:13:44.400
<v Speaker 4>as I mentioned earlier, is face swapping, where you've got

0:13:44.600 --> 0:13:49.680
<v Speaker 4>an individual who's nude performing some sex act and their

0:13:49.720 --> 0:13:52.760
<v Speaker 4>face has been replaced with the face of somebody like

0:13:52.960 --> 0:13:55.760
<v Speaker 4>miss Chew, who you alluded to a couple of minutes back.

0:13:56.000 --> 0:13:59.880
<v Speaker 4>So that's one class of techniques. Another is to use

0:14:00.800 --> 0:14:07.240
<v Speaker 4>massive databases of training data of existing nude individuals to

0:14:07.360 --> 0:14:10.800
<v Speaker 4>create and generate new nudes of people who are hyper

0:14:10.840 --> 0:14:16.120
<v Speaker 4>realistic who are performing acts which the real individual portrayed

0:14:16.480 --> 0:14:19.600
<v Speaker 4>never performed. So that's another class of techniques, and I

0:14:19.600 --> 0:14:21.840
<v Speaker 4>don't want to go into the gory details of how

0:14:21.840 --> 0:14:24.640
<v Speaker 4>that's done, but there are a number of techniques. These

0:14:24.640 --> 0:14:28.480
<v Speaker 4>techniques have evolved better and better every single year, and

0:14:28.520 --> 0:14:32.320
<v Speaker 4>so as a consequence, what we see is really a

0:14:32.440 --> 0:14:36.760
<v Speaker 4>huge number of websites that can disseminate such stuff, not

0:14:36.880 --> 0:14:41.560
<v Speaker 4>just the produced content, but also the tools that are

0:14:41.680 --> 0:14:45.120
<v Speaker 4>used to create and disseminate this stuff. And that's typically

0:14:45.160 --> 0:14:48.479
<v Speaker 4>done on the dark web through some kind of distribution

0:14:48.600 --> 0:14:49.840
<v Speaker 4>network on the dark web.

0:14:50.080 --> 0:14:53.080
<v Speaker 3>And what changed when Elon made it easy to do

0:14:53.160 --> 0:14:55.120
<v Speaker 3>so because for all eyes in the dark web, you

0:14:55.240 --> 0:14:58.280
<v Speaker 3>have to actually make an attempt to go get it right.

0:14:58.320 --> 0:15:00.240
<v Speaker 3>It's like go into a porn store or go going

0:15:00.280 --> 0:15:02.560
<v Speaker 3>to a porn movie or something back in the old days.

0:15:03.080 --> 0:15:05.160
<v Speaker 3>Talk a little bit about how that changed.

0:15:05.440 --> 0:15:09.440
<v Speaker 4>Well, I can't speak specifically to Elon and what he did,

0:15:09.560 --> 0:15:12.840
<v Speaker 4>but what I can say is that the more these

0:15:12.920 --> 0:15:19.239
<v Speaker 4>tools are available through normal channels, the more the audience

0:15:19.680 --> 0:15:23.360
<v Speaker 4>that gets access to them. And as soon as you

0:15:23.400 --> 0:15:26.720
<v Speaker 4>have a big audience, you have a bigger audience of

0:15:26.720 --> 0:15:32.840
<v Speaker 4>bad guys or as Carrie mentioned, just teenagers who don't

0:15:32.840 --> 0:15:37.160
<v Speaker 4>know better and haven't been taught better, possibly because their

0:15:37.200 --> 0:15:41.280
<v Speaker 4>teachers and their parents, however decent they may be, were

0:15:41.320 --> 0:15:44.080
<v Speaker 4>not aware of the threat, and by the time they

0:15:44.120 --> 0:15:46.120
<v Speaker 4>figured out it was too late.

0:15:46.720 --> 0:15:48.720
<v Speaker 3>Deep fakes and newifi apps are also a big problem

0:15:48.760 --> 0:15:51.720
<v Speaker 3>in schools. As we mentioned earlier, recent survey of roughly

0:15:51.720 --> 0:15:54.040
<v Speaker 3>five hundred and fifty US teens found that over half

0:15:54.320 --> 0:15:57.320
<v Speaker 3>had created at least one image using neuification tools, while

0:15:57.320 --> 0:16:00.120
<v Speaker 3>a third of the teens reported having their image created

0:16:00.120 --> 0:16:03.040
<v Speaker 3>and shared non consentually. I'm sure the numbers are higher. Actually,

0:16:03.400 --> 0:16:06.680
<v Speaker 3>Sites like mister deep Figs and apps like groc as

0:16:06.760 --> 0:16:09.880
<v Speaker 3>we know to have normalized creating nons consensual in some

0:16:09.960 --> 0:16:13.600
<v Speaker 3>cases violent sexual imagery, often of women. Laurie talk about

0:16:13.600 --> 0:16:15.480
<v Speaker 3>the consequences for young people.

0:16:15.760 --> 0:16:19.240
<v Speaker 2>Yeah, you know, I'll start. I always like to say,

0:16:19.280 --> 0:16:20.960
<v Speaker 2>like to try to make it a little bit personal.

0:16:21.040 --> 0:16:22.920
<v Speaker 2>I'll never forget, and I'm going to take it here

0:16:22.920 --> 0:16:25.240
<v Speaker 2>and I promise I'll get to young folks, but I'll

0:16:25.320 --> 0:16:27.520
<v Speaker 2>never forget. I think it was twenty twenty three when

0:16:27.560 --> 0:16:30.320
<v Speaker 2>a group of technologists approached me and say, can we

0:16:30.600 --> 0:16:32.360
<v Speaker 2>can we have AI attack you and see what it

0:16:32.400 --> 0:16:36.160
<v Speaker 2>could find. And because I am a crazy person, I said, sure,

0:16:36.320 --> 0:16:39.200
<v Speaker 2>let's see. And what they did was they had AI

0:16:39.320 --> 0:16:41.760
<v Speaker 2>come up with an attack against me. And it took

0:16:41.800 --> 0:16:45.000
<v Speaker 2>some true things, right, like I've interviewed Mark Zuckerberg before,

0:16:45.120 --> 0:16:46.720
<v Speaker 2>or I'm a technology journalist.

0:16:47.080 --> 0:16:49.560
<v Speaker 1>And then it ended up in front of an audience.

0:16:49.800 --> 0:16:52.600
<v Speaker 2>It ended up creating the created sexually explicit deep fix

0:16:52.600 --> 0:16:55.880
<v Speaker 2>of me. And I'll never forget because this I had consented.

0:16:55.920 --> 0:16:57.320
<v Speaker 2>I didn't realize they were going to do that, but

0:16:57.360 --> 0:16:59.880
<v Speaker 2>I had consented to this demo, and I'll never forget

0:17:00.080 --> 0:17:03.200
<v Speaker 2>looking at the audience look at these sexually explicit deep

0:17:03.200 --> 0:17:05.760
<v Speaker 2>fakes of me, and I could, as a person who

0:17:05.960 --> 0:17:08.760
<v Speaker 2>generally formed sentences for a living, I stopped being able

0:17:08.760 --> 0:17:09.600
<v Speaker 2>to form sentences.

0:17:09.760 --> 0:17:12.240
<v Speaker 1>I felt shame, I felt humiliation.

0:17:12.640 --> 0:17:14.840
<v Speaker 2>I felt like the world had seen me naked, to

0:17:14.880 --> 0:17:17.760
<v Speaker 2>be quite honest with you, and then I thought to myself, well,

0:17:17.760 --> 0:17:20.320
<v Speaker 2>if that's how I feel, and I know what this

0:17:20.440 --> 0:17:23.440
<v Speaker 2>technology is, imagine what that's going to be like for

0:17:23.560 --> 0:17:26.480
<v Speaker 2>an eleven year old girl, you know. And I think

0:17:26.560 --> 0:17:28.960
<v Speaker 2>that's what we started hearing. So I've spoken to a

0:17:28.960 --> 0:17:31.800
<v Speaker 2>lot of parents, a lot of teenagers, who've had this

0:17:31.880 --> 0:17:34.520
<v Speaker 2>happen to them. I spoke to one woman who said,

0:17:34.640 --> 0:17:37.000
<v Speaker 2>and you know, she's in her early twenties. She said

0:17:37.040 --> 0:17:39.159
<v Speaker 2>she after this was happening, and she went out and

0:17:39.160 --> 0:17:41.040
<v Speaker 2>she tried to get this stopped, and people said, there's

0:17:41.040 --> 0:17:43.520
<v Speaker 2>nothing you can do. They blamed her, very similar to

0:17:43.560 --> 0:17:45.840
<v Speaker 2>what Carrie is talking about. Back when we were talking

0:17:45.880 --> 0:17:49.159
<v Speaker 2>about non consensual pornography, people didn't quite understand it. She

0:17:49.240 --> 0:17:51.560
<v Speaker 2>told me she walked to the roof of her building

0:17:51.640 --> 0:17:54.399
<v Speaker 2>and she considered jumping off. And you talk to a

0:17:54.440 --> 0:17:57.760
<v Speaker 2>politician who describes this as digital rape. And so for me,

0:17:57.880 --> 0:17:59.879
<v Speaker 2>it was really confusing because I was looking at this

0:18:00.040 --> 0:18:04.080
<v Speaker 2>website that for me, this website represented this dystopian era

0:18:04.160 --> 0:18:07.320
<v Speaker 2>where we lose consent, in an ai era where you know,

0:18:07.400 --> 0:18:10.359
<v Speaker 2>our most intimate qualities can be weaponized against us. And

0:18:10.359 --> 0:18:13.359
<v Speaker 2>then externally we were hearing but it's not real. It

0:18:13.359 --> 0:18:17.720
<v Speaker 2>doesn't matter when it really impacts the way a young person,

0:18:18.359 --> 0:18:20.560
<v Speaker 2>any person walks through the world. We spoke to a

0:18:20.600 --> 0:18:23.479
<v Speaker 2>woman named Mollie who was a mother who had this

0:18:23.680 --> 0:18:27.639
<v Speaker 2>her her husband's best friend. Deep faked the whole neighborhood,

0:18:27.720 --> 0:18:32.120
<v Speaker 2>eighty women and the community and created it's sexually explicit

0:18:32.119 --> 0:18:35.080
<v Speaker 2>deep fakes, and she said it impacted how she would

0:18:35.119 --> 0:18:37.280
<v Speaker 2>go on the bus and look at people and wonder

0:18:37.680 --> 0:18:39.679
<v Speaker 2>if they'd seen this, and she had to go on

0:18:39.760 --> 0:18:42.320
<v Speaker 2>a she said, a porn hunt for her own face

0:18:42.480 --> 0:18:45.960
<v Speaker 2>on the web. And I think it really impacts from

0:18:46.000 --> 0:18:48.320
<v Speaker 2>a human perspective, how you walk through the world and

0:18:48.320 --> 0:18:49.120
<v Speaker 2>how you hold.

0:18:48.880 --> 0:18:51.520
<v Speaker 1>Yourself, and it can have reputational harm.

0:18:51.880 --> 0:18:54.720
<v Speaker 3>Right, So, Carrie, your law firm in your or ensues

0:18:54.840 --> 0:18:58.800
<v Speaker 3>quote psychos pervs, trolls and toxic tech. In January, you're

0:18:58.800 --> 0:19:01.399
<v Speaker 3>from file the lawsuit against say On behalf of Ashley

0:19:01.440 --> 0:19:04.639
<v Speaker 3>Saint Clair, a conservative influencer who had a child with

0:19:04.680 --> 0:19:07.400
<v Speaker 3>Elon Musk. She alleges that his company's Chapbot was used

0:19:07.400 --> 0:19:11.240
<v Speaker 3>to create and disseminate sexually explicit, non consensual images of her.

0:19:11.640 --> 0:19:14.479
<v Speaker 3>Talk about the lawsuit and how is it representative of

0:19:14.600 --> 0:19:16.960
<v Speaker 3>bigger legal issues around deep big porn.

0:19:17.200 --> 0:19:21.399
<v Speaker 5>When Grock suddenly had this new capability of being able

0:19:21.440 --> 0:19:26.600
<v Speaker 5>to create neutification images, this was the first time in

0:19:26.760 --> 0:19:31.520
<v Speaker 5>history that a newdifying tool was integrated into a widely

0:19:31.600 --> 0:19:36.000
<v Speaker 5>used social media platform, which combined the ability to just

0:19:36.480 --> 0:19:42.400
<v Speaker 5>generate an image with the widespread dissemination network, and so

0:19:43.080 --> 0:19:46.160
<v Speaker 5>an image could just be viewable by millions and millions

0:19:46.200 --> 0:19:49.600
<v Speaker 5>of people. But also, I mean, you know, x's for

0:19:49.720 --> 0:19:53.159
<v Speaker 5>people thirteen and up, and so children were also then

0:19:53.359 --> 0:19:58.040
<v Speaker 5>consumers of the images, which is illegal. And we're also

0:19:59.200 --> 0:20:01.119
<v Speaker 5>there were children that were the victims of it. And

0:20:01.359 --> 0:20:04.800
<v Speaker 5>the images posted were not just people in bikinis, but

0:20:05.080 --> 0:20:09.159
<v Speaker 5>GROC would pose them in sexual positions. It would respond

0:20:09.240 --> 0:20:12.879
<v Speaker 5>to prompts telling them to drench their body and blood

0:20:12.920 --> 0:20:17.920
<v Speaker 5>and seamen like fluids, whold suggestive props, and just really

0:20:18.040 --> 0:20:22.760
<v Speaker 5>vulgar and explicit images, and it seemed very clear that

0:20:22.800 --> 0:20:28.120
<v Speaker 5>this was an intended rollout a product that was known

0:20:28.160 --> 0:20:30.960
<v Speaker 5>to be harmful. They were like something like one point

0:20:30.960 --> 0:20:34.800
<v Speaker 5>eight million posts of women and children being deep faked,

0:20:35.160 --> 0:20:37.879
<v Speaker 5>which accounted for like forty one percent of all the

0:20:38.359 --> 0:20:42.480
<v Speaker 5>images posted during this nine day period. So Ashley Saint

0:20:42.480 --> 0:20:46.520
<v Speaker 5>Clair is somebody who is really demonized on the X

0:20:46.640 --> 0:20:49.440
<v Speaker 5>platform because of some of the hostile things that Elon

0:20:49.520 --> 0:20:52.119
<v Speaker 5>Musk has said about her, was a huge target of

0:20:52.160 --> 0:20:54.400
<v Speaker 5>the deep fakes, and there were just so many images

0:20:54.440 --> 0:20:59.720
<v Speaker 5>that GROC generated and published on her own page, and

0:20:59.760 --> 0:21:03.159
<v Speaker 5>so we sued them. We first got you attempted to

0:21:03.160 --> 0:21:07.640
<v Speaker 5>get a temporary restraining order to stop Grock from deepfaking her,

0:21:08.200 --> 0:21:11.600
<v Speaker 5>and then we sued them for defective design and other

0:21:11.960 --> 0:21:15.879
<v Speaker 5>sort of product liability types of lawsuits, negligence, and the

0:21:15.920 --> 0:21:17.920
<v Speaker 5>claim that I like the best is that we sue

0:21:17.920 --> 0:21:19.360
<v Speaker 5>them for public nuisance.

0:21:19.800 --> 0:21:22.359
<v Speaker 3>There's just no reasonable.

0:21:21.920 --> 0:21:26.080
<v Speaker 5>Safe use of this product but to harass people publicly.

0:21:26.200 --> 0:21:27.760
<v Speaker 3>So that's where it stands now.

0:21:27.840 --> 0:21:33.000
<v Speaker 5>And yeah, well, we had to give XAI statutory notice

0:21:33.000 --> 0:21:37.159
<v Speaker 5>that we were filing for a temporary restraining order, and

0:21:37.520 --> 0:21:42.280
<v Speaker 5>they immediately sued Ashley in Texas, claiming that she had

0:21:42.320 --> 0:21:46.560
<v Speaker 5>breached the XAI terms of service by even threatening to

0:21:46.600 --> 0:21:49.280
<v Speaker 5>sue them in her home state of New York. So

0:21:49.359 --> 0:21:53.120
<v Speaker 5>now we are litigating the issue in two different venues.

0:21:53.440 --> 0:21:57.280
<v Speaker 5>They're also trying to use the x terms of service

0:21:57.720 --> 0:22:01.800
<v Speaker 5>to get the XAI case transfer to New York, and

0:22:01.920 --> 0:22:05.440
<v Speaker 5>the judge initially last week said that XAI can rely

0:22:05.600 --> 0:22:08.880
<v Speaker 5>on the x terms of service to transfer.

0:22:08.480 --> 0:22:10.879
<v Speaker 3>The case to Texas. You mean to Texas.

0:22:10.960 --> 0:22:13.840
<v Speaker 5>Yeah, so even though x is not a defendant in

0:22:13.880 --> 0:22:17.800
<v Speaker 5>our case, only XAI is, the judge still let them

0:22:17.920 --> 0:22:21.879
<v Speaker 5>borrow from a different Elon company, which I think is

0:22:21.920 --> 0:22:22.600
<v Speaker 5>really scary.

0:22:22.800 --> 0:22:24.000
<v Speaker 3>So is that where it's going.

0:22:24.359 --> 0:22:27.840
<v Speaker 5>Well, we filed an emergency petition for Mandamus to try

0:22:27.840 --> 0:22:30.040
<v Speaker 5>to get the stay held up, but I don't know,

0:22:30.200 --> 0:22:31.400
<v Speaker 5>we don't have the outcome on that.

0:22:33.640 --> 0:22:45.360
<v Speaker 7>We'll be back in a minute.

0:22:47.960 --> 0:22:51.080
<v Speaker 3>So let's get to the tech infrastructure that supports the

0:22:51.119 --> 0:22:55.040
<v Speaker 3>creation dissemination of deep figs. Analysis from Media Outlet Indicator

0:22:55.280 --> 0:22:58.720
<v Speaker 3>found that newdify apps often use infrastructure from tech companies vs.

0:22:58.840 --> 0:23:02.920
<v Speaker 3>Explain the relationship between deep fake porn sites, tech companies,

0:23:02.920 --> 0:23:05.679
<v Speaker 3>and the social media platforms and to what extent do

0:23:05.760 --> 0:23:09.120
<v Speaker 3>these sites and newify apps rely on big tech infrastructure.

0:23:09.119 --> 0:23:11.840
<v Speaker 3>This was an issue when I was interviewing people who

0:23:11.880 --> 0:23:15.600
<v Speaker 3>were creating all manner of nonsense around the insurrection on

0:23:15.720 --> 0:23:19.240
<v Speaker 3>January sixth, and companies move quickly to take some of

0:23:19.280 --> 0:23:22.400
<v Speaker 3>it down, but they were clearly the way people got

0:23:22.440 --> 0:23:24.200
<v Speaker 3>to these different sites.

0:23:24.600 --> 0:23:27.679
<v Speaker 4>That's a broad question, Kara, so let me try and

0:23:27.720 --> 0:23:32.640
<v Speaker 4>break it down into pieces. The first is depending if

0:23:32.640 --> 0:23:37.160
<v Speaker 4>you're generating deep fakes at scale. So think the kind

0:23:37.160 --> 0:23:39.880
<v Speaker 4>of thing Laurie was talking about with mister deep fake.

0:23:40.240 --> 0:23:44.159
<v Speaker 4>He's trying to generate thousands, millions of deep fakes, and

0:23:44.240 --> 0:23:49.560
<v Speaker 4>so that requires computational hardware what are called graphical computing

0:23:49.680 --> 0:23:54.320
<v Speaker 4>units or GPUs, which today are pretty scarce and expensive.

0:23:55.040 --> 0:23:59.000
<v Speaker 4>So and that scarcity and price doesn't look like it's

0:23:59.080 --> 0:24:01.960
<v Speaker 4>coming down very much much anytime in the near future.

0:24:02.600 --> 0:24:09.200
<v Speaker 4>Those resources are heavily supported by cloud infrastructure providers. Again,

0:24:09.240 --> 0:24:12.119
<v Speaker 4>I'm not naming any specific companies, but there are a

0:24:12.200 --> 0:24:16.639
<v Speaker 4>number of companies that offer cloud services, so part of

0:24:16.680 --> 0:24:21.800
<v Speaker 4>the onus is placed on them if somebody is trying

0:24:21.840 --> 0:24:25.159
<v Speaker 4>to use their service to generate this kind of stuff

0:24:25.240 --> 0:24:28.080
<v Speaker 4>at scale. If somebody is trying to generate just a

0:24:28.080 --> 0:24:30.879
<v Speaker 4>handful of deep fakes, they're probably not going to figure

0:24:30.880 --> 0:24:34.200
<v Speaker 4>it out. But if somebody is generating millions and millions

0:24:34.240 --> 0:24:36.800
<v Speaker 4>of them, or thousands and thousands of them, they may

0:24:37.240 --> 0:24:39.919
<v Speaker 4>have some idea of what's going on. But that's not

0:24:40.080 --> 0:24:45.199
<v Speaker 4>easy either, because you know, there are many challenges for

0:24:45.320 --> 0:24:48.320
<v Speaker 4>some of these firms in figuring out what's a deep

0:24:48.320 --> 0:24:53.600
<v Speaker 4>fake and what's not, what's consensual, what's not and so

0:24:53.840 --> 0:24:57.200
<v Speaker 4>as the first issue, you know, the bad guys are

0:24:57.200 --> 0:25:01.280
<v Speaker 4>continuously evolving, so that me means that the actors in

0:25:01.359 --> 0:25:05.679
<v Speaker 4>question are changing. The technology used to generate this stuff

0:25:05.720 --> 0:25:10.200
<v Speaker 4>is changing, and it turns out, having studied deep fake

0:25:10.240 --> 0:25:13.960
<v Speaker 4>detectors quite a lot and created some on my own

0:25:14.520 --> 0:25:18.680
<v Speaker 4>with my lab, what we know is that small changes

0:25:19.400 --> 0:25:21.960
<v Speaker 4>can make a big difference. So to give you an example,

0:25:22.119 --> 0:25:25.159
<v Speaker 4>if I'm drinking this cup of coffee and there's a

0:25:25.200 --> 0:25:28.840
<v Speaker 4>deep fake of me with part of my face obscure

0:25:28.920 --> 0:25:32.359
<v Speaker 4>as I'm drinking that cup of coffee, it's harder in

0:25:32.400 --> 0:25:36.960
<v Speaker 4>many cases for many detectors to figure out that video

0:25:37.040 --> 0:25:40.000
<v Speaker 4>clip of me saying something which I probably never said

0:25:40.359 --> 0:25:45.400
<v Speaker 4>is not true. Likewise, the definition carry I'm not speaking

0:25:45.400 --> 0:25:49.679
<v Speaker 4>from a legal perspective, but for a normal person, what

0:25:49.920 --> 0:25:54.760
<v Speaker 4>constitutes an act that sexual in nature can vary a lot.

0:25:55.160 --> 0:25:58.879
<v Speaker 4>A sexualized act versus an actual act of you know,

0:25:59.040 --> 0:26:05.160
<v Speaker 4>some kind of sex very different. Also, the relative positions

0:26:05.200 --> 0:26:08.520
<v Speaker 4>in which a sexual act is carried out, where certain

0:26:08.560 --> 0:26:12.240
<v Speaker 4>body parts may be obscured but other things may be clear.

0:26:12.960 --> 0:26:16.480
<v Speaker 4>All that causes machine learning algorithms to detect deep fakes

0:26:16.720 --> 0:26:20.640
<v Speaker 4>to have a lot of problems. So I think those

0:26:20.680 --> 0:26:24.000
<v Speaker 4>are some of the challenges from a technical perspective in

0:26:24.080 --> 0:26:25.240
<v Speaker 4>figuring these things out.

0:26:25.400 --> 0:26:27.800
<v Speaker 3>So one analyst has said that some deep fake sites

0:26:27.800 --> 0:26:31.000
<v Speaker 3>now offer APIs to people creating non consensual images and

0:26:31.119 --> 0:26:33.639
<v Speaker 3>video generators explain what that means.

0:26:34.320 --> 0:26:38.280
<v Speaker 4>So, when you go to a website, you're typing stuff

0:26:38.320 --> 0:26:42.520
<v Speaker 4>in a prompt or a query, or clicking on some

0:26:42.560 --> 0:26:46.399
<v Speaker 4>buttons and you get a response. An API is an

0:26:46.440 --> 0:26:50.640
<v Speaker 4>application program interface, and it replaces you with a piece

0:26:50.680 --> 0:26:54.000
<v Speaker 4>of computer code, and that piece of computer code is

0:26:54.080 --> 0:26:57.919
<v Speaker 4>interacting in a different kind of language with the server

0:26:58.240 --> 0:27:02.680
<v Speaker 4>behind the website you're going to. And so if somebody

0:27:02.720 --> 0:27:07.200
<v Speaker 4>is using an API to generate deep fakes, that suggests

0:27:07.200 --> 0:27:10.080
<v Speaker 4>that they're generating it at scale. That suggests that they

0:27:10.119 --> 0:27:13.320
<v Speaker 4>have a piece of code that is shipping requests automatically

0:27:13.520 --> 0:27:16.239
<v Speaker 4>through that piece of code to the server, and that

0:27:16.320 --> 0:27:19.560
<v Speaker 4>server is sending back whatever they requested, in this case

0:27:19.960 --> 0:27:23.919
<v Speaker 4>some kind of sexualized image. That suggests somebody who's trying

0:27:23.920 --> 0:27:27.800
<v Speaker 4>to do it at scale, who is, you know, who

0:27:27.800 --> 0:27:30.800
<v Speaker 4>probably has his or her own databases of people they

0:27:30.840 --> 0:27:35.320
<v Speaker 4>want to propose in these situations that they will never

0:27:35.400 --> 0:27:36.560
<v Speaker 4>in to begin with.

0:27:37.400 --> 0:27:40.880
<v Speaker 3>Indicators Analysis also estimates that the new toifier economy may

0:27:40.920 --> 0:27:44.159
<v Speaker 3>be worth up to thirty six million dollars a year. Laurie,

0:27:44.200 --> 0:27:46.639
<v Speaker 3>and you're reporting on mister deep fakes. You noted the

0:27:46.640 --> 0:27:50.560
<v Speaker 3>financial incentives for the community members posting in the forum

0:27:50.720 --> 0:27:53.560
<v Speaker 3>tell us about the incentives and talk about the money

0:27:53.560 --> 0:27:56.520
<v Speaker 3>making Laurie here, and for the platforms that host this

0:27:56.680 --> 0:28:00.560
<v Speaker 3>content too, although for people don't realize. Goo Goal says

0:28:00.600 --> 0:28:03.680
<v Speaker 3>it doesn't allow apps to contain sexual content. Apple says

0:28:03.720 --> 0:28:07.400
<v Speaker 3>the company's apps to or prohibits overtly sexual content. Both

0:28:07.440 --> 0:28:11.400
<v Speaker 3>companies disabled searches for new defying their app stores. We'll

0:28:11.400 --> 0:28:13.399
<v Speaker 3>get to the workarounds in a second, but talk a

0:28:13.400 --> 0:28:16.040
<v Speaker 3>little bit aout the incentives, the financial incentives.

0:28:16.320 --> 0:28:19.040
<v Speaker 2>It's interesting because when we were looking at this, this

0:28:19.240 --> 0:28:21.320
<v Speaker 2>site that wasn't a site that was a platform. It

0:28:21.400 --> 0:28:24.440
<v Speaker 2>was a whole ecosystem that enabled the site to exist,

0:28:24.600 --> 0:28:28.080
<v Speaker 2>and that included platform liability from a lot of these

0:28:28.080 --> 0:28:31.439
<v Speaker 2>different companies. So for example, you know, the incentive was

0:28:31.480 --> 0:28:34.080
<v Speaker 2>you can go on not only can you view you know,

0:28:34.240 --> 0:28:37.119
<v Speaker 2>thousands of videos of sexually explicit deep fakes. If your

0:28:37.160 --> 0:28:39.680
<v Speaker 2>favorite celebrity, you can make your own. And this is

0:28:39.680 --> 0:28:42.400
<v Speaker 2>why we believed mister deep Fakes was so dangerous. He

0:28:42.600 --> 0:28:44.840
<v Speaker 2>was dangerous because he was creating an army of mister

0:28:44.880 --> 0:28:47.440
<v Speaker 2>deep fakes. Because now there are all these creators on there,

0:28:47.760 --> 0:28:49.600
<v Speaker 2>which is kind of pay to play, and there's a

0:28:49.640 --> 0:28:52.280
<v Speaker 2>freemium model, and then there's this premium model where you

0:28:52.320 --> 0:28:56.000
<v Speaker 2>can pay to have specialized deep fakes of public figures

0:28:56.000 --> 0:28:58.800
<v Speaker 2>who didn't consent and made and so people were probably

0:28:58.800 --> 0:29:01.240
<v Speaker 2>making more money than the guy behind mister deep fakes

0:29:01.240 --> 0:29:03.680
<v Speaker 2>off of this platform. It was creating a whole ecosystem.

0:29:03.680 --> 0:29:06.400
<v Speaker 2>But what else enabled that ecosystem? First, I would go

0:29:06.400 --> 0:29:09.840
<v Speaker 2>from a regulatory standpoint, the laws just hadn't caught up,

0:29:09.880 --> 0:29:13.320
<v Speaker 2>and we saw this with non consentual pornography. Oftentimes tech

0:29:13.360 --> 0:29:17.320
<v Speaker 2>moves quickly when there are the correct guardrails. Women and

0:29:17.520 --> 0:29:19.960
<v Speaker 2>young folks are the first ones to feel that pain.

0:29:20.360 --> 0:29:23.440
<v Speaker 2>But also there were payment providers that were accepting payments

0:29:23.520 --> 0:29:26.760
<v Speaker 2>on this platform, and so it was all of these things,

0:29:26.760 --> 0:29:30.000
<v Speaker 2>and we talk about these small steps that actually helped

0:29:30.120 --> 0:29:33.080
<v Speaker 2>enable this ecosystem to exist. And one of the coolest

0:29:33.120 --> 0:29:35.800
<v Speaker 2>things we saw because I think so many folks would say, well,

0:29:35.800 --> 0:29:37.440
<v Speaker 2>it's a game of whack a mole. You go after

0:29:37.480 --> 0:29:39.560
<v Speaker 2>this guy and then there are all these other ones

0:29:39.600 --> 0:29:41.960
<v Speaker 2>that are going to pop up, But actually what ended

0:29:42.040 --> 0:29:44.520
<v Speaker 2>up happening is we did go after this guy. We

0:29:44.600 --> 0:29:47.000
<v Speaker 2>found out who he was, and I think maybe most

0:29:47.000 --> 0:29:50.520
<v Speaker 2>alarming was he wasn't a technical guy. This is a

0:29:50.560 --> 0:29:54.120
<v Speaker 2>guy who was, i would say, in his thirties, worked

0:29:54.160 --> 0:29:57.800
<v Speaker 2>at a pharmacist at a hospital, helping people, was newly married,

0:29:57.960 --> 0:30:01.040
<v Speaker 2>had a young child, and he had been able to

0:30:01.120 --> 0:30:05.400
<v Speaker 2>create a website that enabled a whole new wave.

0:30:05.200 --> 0:30:07.120
<v Speaker 1>Of deep fake abuse because.

0:30:06.840 --> 0:30:09.960
<v Speaker 2>He had an interest in technology and deep fakes, and

0:30:10.000 --> 0:30:13.240
<v Speaker 2>he saw an opening because when Reddit banned sexually explicit

0:30:13.320 --> 0:30:16.480
<v Speaker 2>deep fakes, we were able to track where his username

0:30:16.840 --> 0:30:20.360
<v Speaker 2>was talking about seeing an opportunity that could be profitable

0:30:20.640 --> 0:30:23.440
<v Speaker 2>and getting other people to build on this. And so

0:30:23.720 --> 0:30:26.880
<v Speaker 2>all of these things together created i would say, an

0:30:26.880 --> 0:30:30.480
<v Speaker 2>ecosystem that allowed it. And all of these things also

0:30:31.080 --> 0:30:33.800
<v Speaker 2>once people started chipping away at this, once the laws

0:30:33.840 --> 0:30:38.480
<v Speaker 2>started slowly but surely catching up, once Google deranked the platform,

0:30:38.600 --> 0:30:41.600
<v Speaker 2>there were all sorts of small steps that actually created

0:30:41.600 --> 0:30:43.760
<v Speaker 2>a friction that made it harder to exist.

0:30:44.000 --> 0:30:46.520
<v Speaker 3>So let's shift to the current laws and regulations surrounding

0:30:46.520 --> 0:30:48.760
<v Speaker 3>deep fake porn, including the Take It Down Act, the

0:30:48.840 --> 0:30:51.000
<v Speaker 3>law that makes it a federal crime to publish non

0:30:51.000 --> 0:30:54.640
<v Speaker 3>consensual explicit deep fakes. The federal government began enforcing it

0:30:54.720 --> 0:30:58.040
<v Speaker 3>in May, Carrie, in your pursuit of tech platform accountability.

0:30:58.040 --> 0:31:00.640
<v Speaker 3>You've cut up against section two thirty. Have many the

0:31:00.720 --> 0:31:03.800
<v Speaker 3>law that protects platforms for being held legally liable for

0:31:03.880 --> 0:31:07.680
<v Speaker 3>content that users post. You mentioned product liability law as

0:31:07.720 --> 0:31:10.800
<v Speaker 3>a way to stop deep fake porn and get justice

0:31:10.800 --> 0:31:14.400
<v Speaker 3>for victims. So talk about using product liability law.

0:31:14.720 --> 0:31:19.480
<v Speaker 5>Section two thirty was intended to basically just immunize platforms

0:31:19.520 --> 0:31:23.600
<v Speaker 5>for content that third parties created and posted. But when

0:31:23.600 --> 0:31:27.920
<v Speaker 5>it comes to deep fakes, especially like in the situation

0:31:28.000 --> 0:31:31.160
<v Speaker 5>of Groc, we're saying that Grock was the one who

0:31:31.560 --> 0:31:35.520
<v Speaker 5>created this material and also published it, and therefore section

0:31:35.560 --> 0:31:39.320
<v Speaker 5>two thirty shouldn't apply at all. But even if it does,

0:31:40.040 --> 0:31:43.600
<v Speaker 5>by suing them under product liability, we're suing them not

0:31:44.080 --> 0:31:49.920
<v Speaker 5>for publishing the content, but for you know, basically harming

0:31:49.960 --> 0:31:54.360
<v Speaker 5>people through its defective design, defective manufacturing.

0:31:54.520 --> 0:31:56.640
<v Speaker 3>These were all foreseeable uses.

0:31:56.680 --> 0:32:01.960
<v Speaker 5>There's no reasonably safe use for new toifying product, and

0:32:02.000 --> 0:32:05.200
<v Speaker 5>so that's just like a second way to overcome section

0:32:05.240 --> 0:32:11.480
<v Speaker 5>two thirty. Now, XAI, they haven't fully exposed what their

0:32:11.520 --> 0:32:13.800
<v Speaker 5>defenses are going to be in our case, but when

0:32:13.840 --> 0:32:17.760
<v Speaker 5>they were opposing the temporary restraining order, they were claiming

0:32:17.840 --> 0:32:20.560
<v Speaker 5>that they wanted. I mean, so far, this would be

0:32:20.640 --> 0:32:23.520
<v Speaker 5>the only AI product that is saying that Section two

0:32:23.560 --> 0:32:27.560
<v Speaker 5>thirty should apply to it and that they're not responsible

0:32:27.880 --> 0:32:32.880
<v Speaker 5>for what people type in as prompts. But even more

0:32:32.880 --> 0:32:35.880
<v Speaker 5>alarming is that they're claiming that they should have the

0:32:35.920 --> 0:32:40.520
<v Speaker 5>First Amendment protections and that rock should basically be protected

0:32:40.920 --> 0:32:45.160
<v Speaker 5>under free speech, which is a pretty alarming idea. You

0:32:45.160 --> 0:32:50.680
<v Speaker 5>know that AI and a chatbot would have free speech rights.

0:32:51.560 --> 0:32:51.960
<v Speaker 3>They don't.

0:32:52.040 --> 0:32:54.000
<v Speaker 5>You know, so far, we're not at a point where

0:32:54.000 --> 0:32:58.200
<v Speaker 5>we're giving constitutional rights to a chatbot. But again, the

0:32:58.360 --> 0:33:01.320
<v Speaker 5>arguments haven't yet hit this the stage where there's been

0:33:01.360 --> 0:33:02.160
<v Speaker 5>fully brief yet.

0:33:02.680 --> 0:33:05.240
<v Speaker 3>So vas talk about workarounds because a lot of these

0:33:05.840 --> 0:33:08.920
<v Speaker 3>these companies are very clever in terms of workarounds. Porn

0:33:08.920 --> 0:33:11.600
<v Speaker 3>people have always been very clever. How easy is it

0:33:11.640 --> 0:33:14.480
<v Speaker 3>to do workarounds here from a technological point of view?

0:33:15.400 --> 0:33:18.160
<v Speaker 4>You know, I was thinking about this when Lori was

0:33:18.200 --> 0:33:22.040
<v Speaker 4>speaking earlier Kara. You know, she talked about a website

0:33:22.320 --> 0:33:25.000
<v Speaker 4>where somebody goes in and says, I want to create

0:33:25.040 --> 0:33:28.600
<v Speaker 4>a deep fake of such and such person doing such

0:33:28.640 --> 0:33:33.080
<v Speaker 4>and such thing. And so that's a classic example off

0:33:33.640 --> 0:33:38.320
<v Speaker 4>a place where that website is using the APIs that

0:33:38.400 --> 0:33:43.800
<v Speaker 4>you asked about earlier to access a cloud provider where

0:33:43.920 --> 0:33:48.720
<v Speaker 4>this code running degenerate the newtification of the sort designed

0:33:48.720 --> 0:33:52.560
<v Speaker 4>by whoever asked the query. But now think about it

0:33:52.600 --> 0:33:57.840
<v Speaker 4>this way. Let's say the platform has the mechanism. Let's

0:33:57.880 --> 0:34:00.440
<v Speaker 4>say they have a perfect mechanism that doesn't exis yes today,

0:34:00.440 --> 0:34:03.520
<v Speaker 4>but suppose they did to detect the deep fake video

0:34:03.720 --> 0:34:08.400
<v Speaker 4>of some form of sexual activity. Then the workarounds for

0:34:08.480 --> 0:34:11.200
<v Speaker 4>the porn guys that are some simple ones. So one

0:34:11.239 --> 0:34:14.880
<v Speaker 4>of the classic methods to generate deep fakes, regardless of

0:34:14.880 --> 0:34:18.840
<v Speaker 4>whether it's for good or bad, is something called stable diffusion.

0:34:19.560 --> 0:34:24.000
<v Speaker 4>So in stable diffusion, what you do is you, let's say,

0:34:24.000 --> 0:34:26.000
<v Speaker 4>throw in an image. I'll use an image rather than

0:34:26.000 --> 0:34:28.480
<v Speaker 4>a video as my example, but you throw in an

0:34:28.480 --> 0:34:32.759
<v Speaker 4>image of somebody and that image gets converted to some

0:34:33.000 --> 0:34:36.720
<v Speaker 4>very coarse representation. You know, think of this as something

0:34:36.760 --> 0:34:40.640
<v Speaker 4>that's a very skeletal version of that image, and then

0:34:41.040 --> 0:34:46.080
<v Speaker 4>what ends up happening is that skeletal version goes through

0:34:46.080 --> 0:34:51.280
<v Speaker 4>the neuification process, but at the skeletal level, so that's

0:34:51.800 --> 0:34:58.399
<v Speaker 4>skeletalized neutified image doesn't look anything like what a real

0:34:58.560 --> 0:35:01.759
<v Speaker 4>nude might look like, and so you can think of

0:35:01.800 --> 0:35:07.799
<v Speaker 4>it as some kind of computationally weird representation of, you know,

0:35:07.880 --> 0:35:11.440
<v Speaker 4>a person who's about to be sexually abused via a

0:35:11.480 --> 0:35:14.880
<v Speaker 4>deep fake. And then they could do all this on platform,

0:35:15.040 --> 0:35:18.520
<v Speaker 4>on a cloud, and then without ever passing the source

0:35:18.520 --> 0:35:22.920
<v Speaker 4>image or the final image, they bring back this course

0:35:23.000 --> 0:35:26.759
<v Speaker 4>representation of the new deified image and perform the last

0:35:26.800 --> 0:35:29.600
<v Speaker 4>step back on their own servers, where it generates the

0:35:29.600 --> 0:35:34.080
<v Speaker 4>final nude. So that's an example of one way in

0:35:34.120 --> 0:35:37.680
<v Speaker 4>which somebody with you know, just limited computer science knowledge,

0:35:38.040 --> 0:35:41.520
<v Speaker 4>who's familiar with the code, who has access to all

0:35:41.600 --> 0:35:45.920
<v Speaker 4>this code, can generate something while avoiding the guardrails of

0:35:46.000 --> 0:35:48.920
<v Speaker 4>some of the platforms, even if those guardrails are close

0:35:49.000 --> 0:35:49.560
<v Speaker 4>to perfect.

0:35:51.360 --> 0:36:06.799
<v Speaker 3>We'll be back in a minute. The Defiance Act, which

0:36:06.840 --> 0:36:09.560
<v Speaker 3>is aimed at allowing victims of non consensual deep figs

0:36:09.560 --> 0:36:11.719
<v Speaker 3>to suit for damages, passed in the Senate, and it's

0:36:11.760 --> 0:36:14.880
<v Speaker 3>currently stalled in the House. Even though some states have

0:36:14.960 --> 0:36:19.839
<v Speaker 3>criminalized deep figs, enforcement remains low. Laurie, What extent are

0:36:19.840 --> 0:36:22.640
<v Speaker 3>these laws effective when it comes to stopping or at

0:36:22.680 --> 0:36:26.120
<v Speaker 3>least curbing the spread. It's happened before with credit card

0:36:26.120 --> 0:36:29.960
<v Speaker 3>companies and certain businesses. You know, things tend to slow

0:36:30.040 --> 0:36:32.799
<v Speaker 3>things down. But in this case, how important is it

0:36:32.800 --> 0:36:35.240
<v Speaker 3>to have the Take It Down and the Defiance Act?

0:36:35.760 --> 0:36:38.040
<v Speaker 2>I'll answer, and I'm curious for Carry's response too, but

0:36:38.040 --> 0:36:40.840
<v Speaker 2>I think it's incredibly important. You know, the Take It

0:36:40.920 --> 0:36:44.120
<v Speaker 2>Down Act, you know, force tech companies to pay attention

0:36:44.200 --> 0:36:46.960
<v Speaker 2>and quicker. It gives them forty eight hours to take

0:36:47.000 --> 0:36:51.480
<v Speaker 2>down non consensual abuse. But oftentimes when it comes to victims,

0:36:51.760 --> 0:36:54.719
<v Speaker 2>it's really difficult to find recourse for this, you know,

0:36:54.800 --> 0:36:57.840
<v Speaker 2>And so the Defiance Act would create civil penalties that

0:36:57.840 --> 0:37:01.040
<v Speaker 2>would enable them to actually have the ability to go

0:37:01.120 --> 0:37:03.680
<v Speaker 2>after the folks who have done this and have civil recourse,

0:37:03.719 --> 0:37:06.040
<v Speaker 2>which I think makes a big difference. And when we

0:37:06.160 --> 0:37:08.440
<v Speaker 2>started out on our search for mister Deep Fakes, which was,

0:37:08.560 --> 0:37:11.960
<v Speaker 2>let's just be honest, really a search for asking why

0:37:12.040 --> 0:37:14.759
<v Speaker 2>is this allowed to happen? You know, there weren't a

0:37:14.760 --> 0:37:16.440
<v Speaker 2>lot of state laws, and I will say a lot

0:37:16.440 --> 0:37:18.240
<v Speaker 2>of the state laws and a lot of the victims

0:37:18.239 --> 0:37:22.120
<v Speaker 2>and survivors we spoke to. Unfortunately, the onus in many

0:37:22.160 --> 0:37:25.120
<v Speaker 2>ways becomes on these survivors to speak out about it

0:37:25.160 --> 0:37:28.000
<v Speaker 2>and help change laws, and they did in the state level,

0:37:28.080 --> 0:37:31.080
<v Speaker 2>which you know, the state laws changing all of them,

0:37:31.239 --> 0:37:34.960
<v Speaker 2>varying in scope. Civil criminal penalties actually were where a

0:37:35.000 --> 0:37:38.080
<v Speaker 2>lot of these different victims initially had recourse. Which is

0:37:38.080 --> 0:37:40.960
<v Speaker 2>also scary to watch as folks go after regulation at

0:37:40.960 --> 0:37:43.680
<v Speaker 2>the state level because oftentimes the federal government is just

0:37:43.719 --> 0:37:45.920
<v Speaker 2>slow to move. I would say, I hope that the

0:37:45.920 --> 0:37:49.239
<v Speaker 2>Defiance Act passes, it can get through. I think this

0:37:49.280 --> 0:37:51.960
<v Speaker 2>is a bipartisan issue because it's really a child safety

0:37:52.000 --> 0:37:53.000
<v Speaker 2>issue as well.

0:37:53.120 --> 0:37:54.719
<v Speaker 3>Right, which is how they're saying it. And so, by

0:37:54.719 --> 0:37:57.160
<v Speaker 3>the way, Wired investigation found that GROC is still being

0:37:57.239 --> 0:38:01.560
<v Speaker 3>used for non consensual explicit content, this despite regulatory scrutiny

0:38:01.560 --> 0:38:05.160
<v Speaker 3>and multiple lawsuits, and after Musk's XAI said it would

0:38:05.200 --> 0:38:08.839
<v Speaker 3>add restrictions in response to an NBC investigation, the x

0:38:08.880 --> 0:38:12.000
<v Speaker 3>safety account post, do we strictly prohibit users from generating

0:38:12.080 --> 0:38:15.240
<v Speaker 3>non consensual explicit deep fakes and from using our tools

0:38:15.239 --> 0:38:19.680
<v Speaker 3>to undress real people? Explain how some of these legal

0:38:19.760 --> 0:38:23.000
<v Speaker 3>dynamics apply to XAI in that regard. Since you're in

0:38:23.040 --> 0:38:25.800
<v Speaker 3>a lawsuit with them, when it keeps showing up that

0:38:25.840 --> 0:38:29.240
<v Speaker 3>they're doing the same thing, you would accuse them of Well.

0:38:29.040 --> 0:38:31.520
<v Speaker 5>I mean that statement that they say that it's against

0:38:31.600 --> 0:38:35.640
<v Speaker 5>their regulations and they're basically passing the blame to users.

0:38:36.080 --> 0:38:39.120
<v Speaker 5>There's nothing in that statement that suggests that they are

0:38:39.800 --> 0:38:44.040
<v Speaker 5>modifying their product so that deep fakes are no longer

0:38:44.520 --> 0:38:47.560
<v Speaker 5>easy to create. And the problem with things like to

0:38:47.640 --> 0:38:50.080
<v Speaker 5>take it down act it's great, you know, suddenly now

0:38:50.239 --> 0:38:52.759
<v Speaker 5>there's a forty eight hour window of time that tech

0:38:52.760 --> 0:38:57.400
<v Speaker 5>platforms are supposed to remove illegal content, But it has

0:38:57.440 --> 0:39:02.440
<v Speaker 5>no teeth because victims can't enforce it. Victims can't soothe

0:39:02.520 --> 0:39:05.440
<v Speaker 5>the platform and say, hey, you know, it's now been

0:39:05.560 --> 0:39:08.160
<v Speaker 5>forty eight days and that content is still up. So

0:39:08.360 --> 0:39:11.520
<v Speaker 5>the problem with all regulations and all these options is

0:39:11.520 --> 0:39:15.359
<v Speaker 5>that the victims cannot enforce them. And so when we're

0:39:15.400 --> 0:39:18.160
<v Speaker 5>talking about who's the bad guy? VS has talked about

0:39:18.200 --> 0:39:21.120
<v Speaker 5>the bad guys in these situations, in my view, the

0:39:21.160 --> 0:39:25.520
<v Speaker 5>worst guys are the platforms. And so until we have

0:39:25.680 --> 0:39:29.000
<v Speaker 5>regulations that put the victims in the driver's seats to

0:39:29.040 --> 0:39:34.080
<v Speaker 5>not just you know, get criminal justice against the individuals,

0:39:34.160 --> 0:39:38.400
<v Speaker 5>but to actually get justice against the platforms, you know,

0:39:38.480 --> 0:39:41.640
<v Speaker 5>I think the only recourse we have is to use

0:39:41.680 --> 0:39:43.879
<v Speaker 5>our courts and make these companies pay.

0:39:44.360 --> 0:39:46.880
<v Speaker 3>So every episode we get a question from an outside expert.

0:39:46.920 --> 0:39:49.160
<v Speaker 3>Here's yours, and you're all going to answer it, So

0:39:49.280 --> 0:39:51.439
<v Speaker 3>let's listen to it first. Hi.

0:39:51.719 --> 0:39:54.320
<v Speaker 8>I'm Georgia Wells, a tech reporter for The Wall Street Journal.

0:39:54.800 --> 0:39:56.799
<v Speaker 8>I just finished a deep dive into the issue of

0:39:56.880 --> 0:40:00.080
<v Speaker 8>teens using deep fake nudes to harass each other, and

0:40:00.120 --> 0:40:02.160
<v Speaker 8>so the big question I would ask is that, now

0:40:02.160 --> 0:40:04.640
<v Speaker 8>that the government has just started enforcing the Take It

0:40:04.680 --> 0:40:08.279
<v Speaker 8>Down Act, how might this change the harassment landscape for

0:40:08.360 --> 0:40:09.200
<v Speaker 8>young people today?

0:40:09.960 --> 0:40:14.640
<v Speaker 3>Thank you, Laura, you go first, then Carrie then VS. Yeah.

0:40:14.719 --> 0:40:17.000
<v Speaker 2>I do agree a little bit with Carrie on saying,

0:40:17.120 --> 0:40:19.399
<v Speaker 2>you know, there's only so much you can actually kind

0:40:19.400 --> 0:40:22.160
<v Speaker 2>of enforce when the onus is on the victim. I

0:40:22.239 --> 0:40:26.480
<v Speaker 2>do think from a cultural standpoint, it is very different

0:40:26.480 --> 0:40:28.480
<v Speaker 2>than it was a couple of years ago of federal

0:40:28.520 --> 0:40:31.640
<v Speaker 2>law Take It Down Act. Again, state laws are at

0:40:31.719 --> 0:40:35.480
<v Speaker 2>least sending this message that the stakes will be high

0:40:35.520 --> 0:40:39.439
<v Speaker 2>and that young boys, young teenagers can get into real

0:40:39.480 --> 0:40:41.719
<v Speaker 2>trouble for doing this and can get charged. I mean,

0:40:42.320 --> 0:40:45.680
<v Speaker 2>I think that at least from a perspective of schools

0:40:45.719 --> 0:40:48.160
<v Speaker 2>being able to speak to students, parents being able to

0:40:48.200 --> 0:40:51.200
<v Speaker 2>speak to their children about this, and having some structure

0:40:51.239 --> 0:40:54.479
<v Speaker 2>will be helpful again, though I'm not sure the Take It.

0:40:54.400 --> 0:40:56.160
<v Speaker 1>Down Act is going to be the thing that's going

0:40:56.200 --> 0:40:57.040
<v Speaker 1>to actually change this.

0:40:57.160 --> 0:40:59.759
<v Speaker 2>I also agree with Kerrie on platform liability in creating

0:40:59.800 --> 0:41:03.959
<v Speaker 2>high stakes for the tech companies and the whole ecosystem

0:41:03.960 --> 0:41:05.520
<v Speaker 2>that enables this to happen.

0:41:05.520 --> 0:41:08.920
<v Speaker 5>Kerrie, So I think they'll get the memo that this

0:41:09.080 --> 0:41:13.319
<v Speaker 5>is illegal. I mean, having criminal laws was really really

0:41:13.360 --> 0:41:17.200
<v Speaker 5>effective when we were dealing with the scourge of image

0:41:17.239 --> 0:41:20.520
<v Speaker 5>asexual abuse. Over the last decade. We went from having

0:41:20.520 --> 0:41:23.680
<v Speaker 5>three states to having fifty states with those laws, and

0:41:23.920 --> 0:41:27.879
<v Speaker 5>my firm saw such an enormous and quick plummet of

0:41:27.880 --> 0:41:32.839
<v Speaker 5>offenders creating images like that. So that's great, But we

0:41:32.880 --> 0:41:37.520
<v Speaker 5>still don't have an apparatus that targets the platforms themselves

0:41:37.840 --> 0:41:43.279
<v Speaker 5>and prevents them from basically monetizing this kind of humiliation.

0:41:43.960 --> 0:41:48.120
<v Speaker 5>And until we have laws that victims can enforce themselves, like,

0:41:48.280 --> 0:41:53.360
<v Speaker 5>it's not really harming they who I consider the true offenders.

0:41:53.000 --> 0:41:56.120
<v Speaker 3>Right, which are the platforms themselves, all right? Vs.

0:41:56.280 --> 0:42:01.360
<v Speaker 4>You know, I have very mixed feelings about criminalizing a

0:42:01.440 --> 0:42:04.600
<v Speaker 4>thirteen year old, all right, But at the same time,

0:42:05.080 --> 0:42:08.080
<v Speaker 4>we need to protect perhaps other thirteen year olds or

0:42:08.160 --> 0:42:11.560
<v Speaker 4>fourteen year olds or ten year olds from the consequences

0:42:12.080 --> 0:42:14.640
<v Speaker 4>that accrue to them because of some thirteen year old

0:42:15.120 --> 0:42:18.879
<v Speaker 4>putting out deep porn of that kid. So I think,

0:42:19.000 --> 0:42:22.160
<v Speaker 4>you know, a first step has got to be something

0:42:22.800 --> 0:42:28.279
<v Speaker 4>around education. A lot of our teachers in schools are

0:42:28.600 --> 0:42:34.839
<v Speaker 4>simply not teaching their kids about deep fakes. And it's

0:42:34.920 --> 0:42:38.280
<v Speaker 4>one thing if kids are using deep fakes or generative

0:42:38.320 --> 0:42:43.040
<v Speaker 4>AI to you know, create reports for their class projects.

0:42:43.040 --> 0:42:44.680
<v Speaker 4>I mean that brings up a whole bunch of other

0:42:44.680 --> 0:42:48.560
<v Speaker 4>ethical issues. But it's another thing if they use deep

0:42:48.600 --> 0:42:54.680
<v Speaker 4>fakes to explicitly cause damage to a kid and one

0:42:54.719 --> 0:42:58.040
<v Speaker 4>of their classmates. So I think the ethics around us

0:42:58.400 --> 0:43:01.600
<v Speaker 4>has got to be clearly or take in class and

0:43:01.840 --> 0:43:04.520
<v Speaker 4>by schools. And some of us schools may have done

0:43:04.520 --> 0:43:08.080
<v Speaker 4>a good job of this, others may not. And you

0:43:08.080 --> 0:43:13.239
<v Speaker 4>know that education and the consequences of those actions in

0:43:13.280 --> 0:43:17.280
<v Speaker 4>a school context are clear. I think it'll be challenging.

0:43:17.560 --> 0:43:20.120
<v Speaker 2>You know, viaz just to say to respond if it's helpful,

0:43:20.160 --> 0:43:22.920
<v Speaker 2>just to respond to that. I think the frustration is

0:43:23.000 --> 0:43:25.920
<v Speaker 2>so many times, I agree, how do we expect thirteen

0:43:26.000 --> 0:43:29.080
<v Speaker 2>year old boys to know that this isn't okay when

0:43:29.080 --> 0:43:32.000
<v Speaker 2>they're getting served up neuification apps on social media.

0:43:32.480 --> 0:43:33.799
<v Speaker 1>You know, I think it's it is.

0:43:34.120 --> 0:43:36.360
<v Speaker 2>It's an end like two things can be true, and

0:43:36.440 --> 0:43:39.359
<v Speaker 2>so there's a huge gap between education and being able

0:43:39.440 --> 0:43:42.440
<v Speaker 2>to educate folks on this and what the platforms are

0:43:42.480 --> 0:43:45.240
<v Speaker 2>allowing and what they say they're allowing and what's coming through.

0:43:45.360 --> 0:43:47.600
<v Speaker 1>It creates a really difficult environment.

0:43:48.000 --> 0:43:52.400
<v Speaker 4>You know, I worry generally about the broad statement or

0:43:52.480 --> 0:43:57.760
<v Speaker 4>the broad platforms Moniker. I think there's a wide variety

0:43:58.200 --> 0:44:01.680
<v Speaker 4>in what different platforms are doing. Some platforms are going

0:44:01.719 --> 0:44:05.040
<v Speaker 4>to a great amount of trouble, spending a tremendous amount

0:44:05.080 --> 0:44:09.360
<v Speaker 4>of money to try and perform the best innovation possible

0:44:09.760 --> 0:44:12.640
<v Speaker 4>so they can put some of these problems to bed,

0:44:13.160 --> 0:44:16.040
<v Speaker 4>and others are doing much much less. So there's like

0:44:16.080 --> 0:44:20.000
<v Speaker 4>a huge range of what platforms are doing, some doing

0:44:20.040 --> 0:44:23.719
<v Speaker 4>relatively little, some making every effort to try and get there.

0:44:24.280 --> 0:44:28.719
<v Speaker 4>Now I want to add that, you know, again, even

0:44:28.760 --> 0:44:31.840
<v Speaker 4>if platforms who are doing their very best, there's a

0:44:31.840 --> 0:44:35.920
<v Speaker 4>famous saying, due to a statistician at the University of Wisconsin,

0:44:36.719 --> 0:44:40.319
<v Speaker 4>all models are wrong, some are useful, all right, And

0:44:40.719 --> 0:44:44.680
<v Speaker 4>so whatever models are being used by platforms to find

0:44:44.719 --> 0:44:48.160
<v Speaker 4>these deep fakes. They're never going to be perfect, and

0:44:48.239 --> 0:44:51.399
<v Speaker 4>so the question is are they making the effort? Some

0:44:51.440 --> 0:44:54.319
<v Speaker 4>are making what I would consider a very good faith effort.

0:44:54.800 --> 0:44:57.919
<v Speaker 3>Some who is doing a good job.

0:44:58.000 --> 0:45:00.200
<v Speaker 4>I'd prefer to stay out of that, but I don't

0:45:00.200 --> 0:45:01.839
<v Speaker 4>want to talk about specific platforms.

0:45:01.880 --> 0:45:04.960
<v Speaker 3>But yeah, so let's end by looking at what individuals

0:45:04.960 --> 0:45:08.800
<v Speaker 3>can do as deep fake technology continues to evolve. Carrying

0:45:08.840 --> 0:45:11.200
<v Speaker 3>now that the Take It Down Act forces tech companies

0:45:11.239 --> 0:45:13.680
<v Speaker 3>to remove deep fake nudes, how do you request a

0:45:13.800 --> 0:45:16.040
<v Speaker 3>takedown if you're targeted? Well, part of what.

0:45:15.960 --> 0:45:20.520
<v Speaker 5>The Takedown Act required was that there be a flow

0:45:21.239 --> 0:45:26.080
<v Speaker 5>for people to easily request content removal, and so that

0:45:26.200 --> 0:45:29.839
<v Speaker 5>kind of infrastructure I really appreciated that being incorporated into

0:45:29.880 --> 0:45:33.520
<v Speaker 5>the laws. So ideally, you know, a victim within the

0:45:33.880 --> 0:45:37.120
<v Speaker 5>platform itself can now report it, you know, whether it's

0:45:37.120 --> 0:45:41.080
<v Speaker 5>through a URL or if it's like a Instagram or Facebook,

0:45:41.080 --> 0:45:44.600
<v Speaker 5>then the image itself would ideally have a method to

0:45:44.800 --> 0:45:46.600
<v Speaker 5>request removal.

0:45:47.040 --> 0:45:50.680
<v Speaker 3>When digital forensics professor and deep fake expert Hani Farred

0:45:50.840 --> 0:45:53.560
<v Speaker 3>was on the podcast in January, he expressed concerns that

0:45:53.680 --> 0:45:56.920
<v Speaker 3>agentic AI will supercharge the creation of deep figs in

0:45:56.960 --> 0:45:59.400
<v Speaker 3>the coming months vs. How do we make sure the

0:45:59.440 --> 0:46:02.840
<v Speaker 3>new laws policies don't block positive uses in these tools.

0:46:03.320 --> 0:46:05.160
<v Speaker 4>So I think, again, if we don't want to block

0:46:05.400 --> 0:46:09.400
<v Speaker 4>the technology, we want the technology to evolve. Because again,

0:46:10.160 --> 0:46:14.000
<v Speaker 4>as Laurie said much earlier in the podcast today, you know,

0:46:14.560 --> 0:46:18.920
<v Speaker 4>deep fakes do have the capacity to help do things

0:46:19.040 --> 0:46:22.080
<v Speaker 4>like better understand disease. I'm not going to call this

0:46:22.160 --> 0:46:25.680
<v Speaker 4>a deep fake, but the generative AI is being used

0:46:25.680 --> 0:46:29.239
<v Speaker 4>to create new kinds of proteins that will help cure diseases.

0:46:29.600 --> 0:46:33.040
<v Speaker 4>Generative AI is being used to create materials that will

0:46:33.080 --> 0:46:35.799
<v Speaker 4>degrade and perform in ways that we would like, that

0:46:35.840 --> 0:46:38.520
<v Speaker 4>would protect the environment. And more so, I think what

0:46:38.560 --> 0:46:41.759
<v Speaker 4>we have to do is to look at intent. You know,

0:46:42.239 --> 0:46:46.160
<v Speaker 4>what is the intent of the individual who is creating

0:46:46.200 --> 0:46:49.960
<v Speaker 4>a deep fake. And in the case of the thirteen

0:46:50.040 --> 0:46:52.800
<v Speaker 4>year old boy or fourteen year old boy who's creating

0:46:53.000 --> 0:46:56.880
<v Speaker 4>deep fakes of one of his classmates because he's pissed

0:46:56.920 --> 0:47:03.080
<v Speaker 4>off about something, I think they're there's intent. Whereas if

0:47:03.080 --> 0:47:05.880
<v Speaker 4>somebody you know, there was this guy who created a

0:47:05.920 --> 0:47:09.360
<v Speaker 4>deep fake of the Pope but Francis in puffy robes

0:47:09.400 --> 0:47:12.839
<v Speaker 4>and a very dapper looking version of him. You know

0:47:13.160 --> 0:47:16.440
<v Speaker 4>that I think is something that's laughable, and I find

0:47:16.480 --> 0:47:19.719
<v Speaker 4>that at least I find that somewhat entertaining. I doubt

0:47:19.800 --> 0:47:22.880
<v Speaker 4>very much of the Pope was, you know, horribly offended.

0:47:23.400 --> 0:47:26.000
<v Speaker 4>And so, you know, I think we want to allow

0:47:26.120 --> 0:47:31.319
<v Speaker 4>artistic freedom and expression scientific inquiry at the same time

0:47:31.560 --> 0:47:35.279
<v Speaker 4>regulate so that it's the uses that are banned or

0:47:35.840 --> 0:47:36.520
<v Speaker 4>that lead.

0:47:36.360 --> 0:47:39.759
<v Speaker 3>To the individual intent one. Yeah, Laurie, In the end,

0:47:39.800 --> 0:47:42.799
<v Speaker 3>you track down mister deep fakes, what ultimately brought down

0:47:42.800 --> 0:47:44.200
<v Speaker 3>the site? And where is he now?

0:47:44.600 --> 0:47:47.400
<v Speaker 1>Great question? Where is he now? I would love to know.

0:47:47.640 --> 0:47:50.920
<v Speaker 2>So eventually we did track him down with the help

0:47:51.040 --> 0:47:54.840
<v Speaker 2>of you know, investigative journalists the internet and some of

0:47:54.840 --> 0:47:57.600
<v Speaker 2>our cybersecurity friends who did some pro bono work for

0:47:57.680 --> 0:47:59.799
<v Speaker 2>us because they have kids and they viewed this as

0:47:59.800 --> 0:48:02.439
<v Speaker 2>a threat. And we were able to track him down,

0:48:02.719 --> 0:48:05.040
<v Speaker 2>and you know, Carol, like I'll never forget confronting him

0:48:05.120 --> 0:48:07.720
<v Speaker 2>on his way into the office. And I do remember

0:48:07.760 --> 0:48:11.040
<v Speaker 2>a Carrie who I've known because I've covered nonconsensual pornography before.

0:48:11.080 --> 0:48:13.759
<v Speaker 2>I called her before going to his parents' home to

0:48:13.800 --> 0:48:16.480
<v Speaker 2>make sure that we did everything correct, just in case,

0:48:16.719 --> 0:48:19.560
<v Speaker 2>you know, we didn't want to create any legal liabilities.

0:48:19.920 --> 0:48:22.360
<v Speaker 2>But I will never forget him looking at me. He

0:48:22.440 --> 0:48:25.359
<v Speaker 2>knew who I was with so much disdain. And I

0:48:25.400 --> 0:48:27.759
<v Speaker 2>know this is a delusional part of me, but there

0:48:27.840 --> 0:48:29.680
<v Speaker 2>was a part of me that thought maybe we would

0:48:29.719 --> 0:48:32.000
<v Speaker 2>have tracked this guy down and we could show up

0:48:32.040 --> 0:48:35.160
<v Speaker 2>and say, do you understand the real harm this site

0:48:35.200 --> 0:48:39.600
<v Speaker 2>is causing? You're a father, And I remember he walked

0:48:39.680 --> 0:48:42.200
<v Speaker 2>right into the hospital. He wouldn't say a word to me,

0:48:42.239 --> 0:48:44.960
<v Speaker 2>and the doors closed, and I remember thinking at that point, well,

0:48:44.960 --> 0:48:46.800
<v Speaker 2>we're gonna have to go talk about this to everyone

0:48:46.880 --> 0:48:49.480
<v Speaker 2>who will listen to us, and that has to change

0:48:49.480 --> 0:48:51.520
<v Speaker 2>because we have to change culture as a part of this.

0:48:51.520 --> 0:48:53.040
<v Speaker 1>This can't be where the story ends.

0:48:53.080 --> 0:48:55.799
<v Speaker 2>He can't just take himself off social media, which he

0:48:55.880 --> 0:48:58.839
<v Speaker 2>did and continue running this site. And so a lot

0:48:58.920 --> 0:49:01.520
<v Speaker 2>of things happened, which I think, think we're really exciting.

0:49:01.600 --> 0:49:03.560
<v Speaker 2>I mean we I think were one of the first

0:49:03.719 --> 0:49:07.240
<v Speaker 2>to show up and threaten his anonymity, so you cannot

0:49:07.360 --> 0:49:09.839
<v Speaker 2>do this type of abuse and remain anonymous.

0:49:09.840 --> 0:49:12.160
<v Speaker 1>So that was friction. That exposure was friction.

0:49:12.680 --> 0:49:15.160
<v Speaker 2>A laws started changing when in the UK they started

0:49:15.200 --> 0:49:17.920
<v Speaker 2>talking about different types of laws that would make it

0:49:17.960 --> 0:49:20.319
<v Speaker 2>harder for folks to access mister deep Fix. They caught

0:49:20.320 --> 0:49:23.080
<v Speaker 2>off access to folks in the UK. That created friction.

0:49:23.760 --> 0:49:26.439
<v Speaker 2>A very valuable server went down for them to take

0:49:26.480 --> 0:49:30.680
<v Speaker 2>it down. At past the CBC published their investigation as well.

0:49:31.040 --> 0:49:34.400
<v Speaker 2>Mister deep Fix finally went offline for good. And I

0:49:34.440 --> 0:49:37.200
<v Speaker 2>think it was such a great lesson in all of

0:49:37.239 --> 0:49:41.440
<v Speaker 2>these players that actually impact how we build out a

0:49:41.480 --> 0:49:44.840
<v Speaker 2>better future. And the headline is friction. It just created

0:49:44.920 --> 0:49:48.200
<v Speaker 2>more and more friction. Now what's happened to him, I

0:49:48.239 --> 0:49:51.120
<v Speaker 2>think is actually is just as important of a headline,

0:49:51.120 --> 0:49:53.919
<v Speaker 2>which is nothing really. I mean, he lost his job

0:49:54.440 --> 0:49:58.800
<v Speaker 2>at the pharmacy. But we've talked to many folks law enforcement.

0:49:58.840 --> 0:50:02.040
<v Speaker 2>I actually spoke to a couple different folks in Canada.

0:50:02.040 --> 0:50:05.280
<v Speaker 2>From the law enforcement perspective, there's no open investigations into

0:50:05.600 --> 0:50:08.839
<v Speaker 2>David Doe. That's his name, and there's thoughts that maybe

0:50:08.880 --> 0:50:11.520
<v Speaker 2>he's left the country. But that's it and I think

0:50:11.560 --> 0:50:14.839
<v Speaker 2>that in and of itself tells its own story absolutely.

0:50:14.960 --> 0:50:17.840
<v Speaker 3>So the last question dve Forne obviously isn't going away.

0:50:18.000 --> 0:50:21.200
<v Speaker 3>So what's the best possible outcome each of you, what

0:50:21.280 --> 0:50:23.799
<v Speaker 3>is your most optimistic case for how it pans out

0:50:23.800 --> 0:50:26.480
<v Speaker 3>in the next five years and how we get there. Carrie,

0:50:26.600 --> 0:50:29.960
<v Speaker 3>you go first, and then VS and Laurie you get

0:50:29.960 --> 0:50:31.040
<v Speaker 3>the last word. Carrie.

0:50:31.160 --> 0:50:33.560
<v Speaker 5>Okay, Well it should probably come as no shock too

0:50:33.719 --> 0:50:36.920
<v Speaker 5>that my best case outcome is that the companies that

0:50:37.480 --> 0:50:42.319
<v Speaker 5>unleashed these harms and profited significantly from them should have

0:50:42.400 --> 0:50:46.600
<v Speaker 5>to pay and there should be disgorgement and compensation for

0:50:46.680 --> 0:50:50.279
<v Speaker 5>all the victims, and it should be financially penalizing for

0:50:50.360 --> 0:50:51.120
<v Speaker 5>these companies.

0:50:51.760 --> 0:50:54.520
<v Speaker 3>So that's would be your best And is that a

0:50:54.760 --> 0:50:57.600
<v Speaker 3>possible outcome? Do you feel like that's a possible outcome.

0:50:58.760 --> 0:51:01.919
<v Speaker 5>I mean, we're on that road with litigation. We'll see

0:51:01.920 --> 0:51:04.920
<v Speaker 5>how it works. I also, you know, I do condemn

0:51:05.520 --> 0:51:08.759
<v Speaker 5>the platforms like the App Store and Google Play who

0:51:09.360 --> 0:51:15.080
<v Speaker 5>for many many years were offering and selling newdifying apps

0:51:15.440 --> 0:51:18.719
<v Speaker 5>and they just kind of were ostriches with their head

0:51:18.719 --> 0:51:21.399
<v Speaker 5>in the sand, pretending like they didn't think these things

0:51:21.400 --> 0:51:25.239
<v Speaker 5>were harmful. I'm really sick and tired of platforms not

0:51:25.400 --> 0:51:29.040
<v Speaker 5>doing anything until something becomes actually criminal. You know, they

0:51:29.080 --> 0:51:33.280
<v Speaker 5>deny that it's harmful until there's an actual criminal law VS.

0:51:34.160 --> 0:51:36.920
<v Speaker 4>You know, that's what I would like in place, some

0:51:37.120 --> 0:51:40.880
<v Speaker 4>kind of criminal law that says, well, look, here are

0:51:40.920 --> 0:51:45.480
<v Speaker 4>the kinds of behaviors, and even intended behaviors are intended

0:51:45.520 --> 0:51:49.040
<v Speaker 4>outcomes that should be criminalized. Here are the penalties that

0:51:49.080 --> 0:51:52.520
<v Speaker 4>you will face if you engage in these acts. And

0:51:52.960 --> 0:51:56.000
<v Speaker 4>this needs to be something that's at least in the

0:51:56.120 --> 0:52:00.200
<v Speaker 4>US nationwide. But more importantly, many of these actors are

0:52:00.239 --> 0:52:03.239
<v Speaker 4>going to be overseas. And so what we need is

0:52:03.719 --> 0:52:07.239
<v Speaker 4>not just US law, but law in various countries and

0:52:07.320 --> 0:52:12.640
<v Speaker 4>some broad international agreements, perhaps enforced by an organization like Interpol,

0:52:13.040 --> 0:52:17.160
<v Speaker 4>which polices things like child grafficking already, they should be

0:52:17.200 --> 0:52:19.920
<v Speaker 4>authorized to do and investigate things like this.

0:52:20.480 --> 0:52:23.160
<v Speaker 3>That's a very good point. That's exactly where it goes

0:52:23.200 --> 0:52:25.920
<v Speaker 3>as broad is what happens is immediately it goes abroad.

0:52:26.239 --> 0:52:29.239
<v Speaker 2>Laurie, last word, Look, I agree with what everyone here

0:52:29.280 --> 0:52:31.200
<v Speaker 2>has said, So I will end with saying, you know,

0:52:31.280 --> 0:52:33.560
<v Speaker 2>my gripe with Silicon Valley, and I have a lot

0:52:33.600 --> 0:52:37.120
<v Speaker 2>of gripes with Silicon Valley, is they're sitting there talking

0:52:37.280 --> 0:52:41.279
<v Speaker 2>about AGI and what happens when AI becomes smarter than us,

0:52:41.680 --> 0:52:44.520
<v Speaker 2>and it doesn't matter, It really doesn't matter, because here's

0:52:44.520 --> 0:52:45.680
<v Speaker 2>what's happening.

0:52:45.280 --> 0:52:46.040
<v Speaker 1>Here on Earth.

0:52:46.480 --> 0:52:49.640
<v Speaker 2>Over here in reality, we have children who are ending

0:52:49.680 --> 0:52:52.680
<v Speaker 2>their lives because of deep fake sex stortion, and people

0:52:52.719 --> 0:52:56.160
<v Speaker 2>are sleeping on it. My hope is that I don't

0:52:56.160 --> 0:52:58.799
<v Speaker 2>have to shout into the ether for years and say

0:52:58.800 --> 0:53:00.600
<v Speaker 2>this is real harm, this is real home. That we

0:53:00.640 --> 0:53:03.799
<v Speaker 2>can get better education, that we can change culture to

0:53:04.040 --> 0:53:08.799
<v Speaker 2>understand that when new technology comes quickly without the correct guardrails,

0:53:09.040 --> 0:53:12.680
<v Speaker 2>the folks who are impacted oftentimes are the ones that

0:53:12.760 --> 0:53:15.360
<v Speaker 2>don't have the voice, and we can make some changes

0:53:15.400 --> 0:53:18.000
<v Speaker 2>from there. No, it doesn't take, to Carrie's point, Oftentimes

0:53:18.040 --> 0:53:20.759
<v Speaker 2>it takes criminal law if I'm putting sitting in my seat.

0:53:20.840 --> 0:53:24.279
<v Speaker 2>Oftentimes it takes public pressure and investigations. I'd love to

0:53:24.320 --> 0:53:26.520
<v Speaker 2>get to a world where Silicon Valley maybe has some

0:53:26.600 --> 0:53:29.800
<v Speaker 2>different folks in it. And this isn't the last thought

0:53:29.920 --> 0:53:32.959
<v Speaker 2>as people are having dinner and talking about AGI after

0:53:33.000 --> 0:53:34.160
<v Speaker 2>taking a cold plunge.

0:53:34.239 --> 0:53:34.640
<v Speaker 1>That's it.

0:53:34.920 --> 0:53:36.520
<v Speaker 3>That's a very good point. But you're going to wait

0:53:36.560 --> 0:53:38.800
<v Speaker 3>a long time for that. Just thank you, But I

0:53:38.880 --> 0:53:45.200
<v Speaker 3>got manifested, go ahead, all right? Sure? Why not? Why not? Anyway?

0:53:45.239 --> 0:53:47.120
<v Speaker 3>I really appreciate it, all three of you. It's a

0:53:47.160 --> 0:53:49.080
<v Speaker 3>really important talk, and we're going to keep coming back

0:53:49.080 --> 0:53:51.480
<v Speaker 3>to it again and again and hopefully see something happen

0:53:51.719 --> 0:53:55.359
<v Speaker 3>by raising awareness of it. For sure. Thank you so much, all.

0:53:55.280 --> 0:53:57.960
<v Speaker 4>Three of you, Thank you, Ker pleasure to be here.

0:53:57.840 --> 0:54:01.600
<v Speaker 3>Thank you so much. Thanks. We've reached out to XAI

0:54:01.640 --> 0:54:06.439
<v Speaker 3>and x for comment. They did not respond. Today's show

0:54:06.480 --> 0:54:10.800
<v Speaker 3>was produced by Michelle Alloy, Katherine Millsop, Madeline LaPlante, Doobie,

0:54:10.920 --> 0:54:15.160
<v Speaker 3>and Kaylin Lynch. Special thanks to Lissa Soap, Anika Robbins

0:54:15.239 --> 0:54:18.920
<v Speaker 3>and Julia Sharp Levine. Our engineers are Fernando Aruda and

0:54:19.000 --> 0:54:22.760
<v Speaker 3>Rick Kwan, and our theme music is by Trackademics. Nishat

0:54:22.840 --> 0:54:26.880
<v Speaker 3>Kerwa is Vox Media's executive producer of podcasts. Go Wherever

0:54:26.960 --> 0:54:29.400
<v Speaker 3>you listen to podcasts, search for On with Kara Swisher

0:54:29.400 --> 0:54:31.919
<v Speaker 3>and hit follow. Thanks for listening to On with Kara

0:54:32.000 --> 0:54:34.800
<v Speaker 3>Swisher from Podium Media, New York Magazine, the vox Media

0:54:34.880 --> 0:54:38.279
<v Speaker 3>Podcast Network and US. We'll be back on Monday with

0:54:38.440 --> 0:54:38.680
<v Speaker 3>more