Latest / The Jon and Marc Podcast / From Section 230 to OpenAI: The New Frontier of Plaintiff Law

Transcript

From Section 230 to OpenAI: The New Frontier of Plaintiff Law · The Jon and Marc Podcast

• 0:00
  The Zane Shamblin case which we filed against ChatGPT was so
• 0:05
  egregious that we actually accused them of manslaughter.
• 0:08
  What we saw was a significant number of our clients were
• 0:12
  utilizing ChatGPT version 3 when 4 Point O came out.
• 0:16
  In many cases had severe downward logical health and in a
• 0:20
  number of cases with their lives.
• 0:30
  Welcome back to the John and Mark Podcast.
• 0:32
  We are here at MTMP at the Win and we have Matt Bergman joining
• 0:38
  us again for the second time in two months.
• 0:41
  Last time we spoke, we were out in California.
• 0:43
  It was the second day of the trial.
• 0:45
  It was I believe the day before Mosseri was testifying.
• 0:50
  So right at the beginning of the trial.
• 0:51
  So much has happened over the past two months.
• 0:54
  We're going to get into what happened during the trial.
• 0:56
  We're going to talk about what it now means and we're going to
• 0:59
  talk about the cases that you're now going to be looking at as
• 1:02
  the first domino has fallen in social media addiction.
• 1:05
  Thanks for coming back and being actually the 2nd guest that we
• 1:08
  have had for the second time. Yeah, you're.
• 1:10
  Second ever recurring guest. OK, Kelsey Stokes was the first.
• 1:15
  We talked to her about hernia mesh last MTMP.
• 1:17
  She came back, talked to us about Olympus scopes yesterday,
• 1:20
  and now you're back. Good to be here.
• 1:22
  By Popular. Demand by popular demand.
• 1:24
  One of them, the most viewed episode of all time of the John
• 1:27
  Moore podcast. So Oh.
• 1:28
  My gosh. We'll send you an award that we
• 1:30
  should probably should have done that.
• 1:32
  Well, it it I guess it happened this year at the end of the
• 1:35
  year, unless you're toppled. We'll see.
• 1:36
  I. I was thinking more of a
• 1:38
  rotating trophy that that if someone else ever gets more
• 1:40
  views, Matt can just mail it to him.
• 1:42
  Let's talk about like, what has what has happened since the
• 1:47
  trial ended? Because that everything
• 1:49
  everybody knows, like, you know, Mark did a fantastic job at
• 1:53
  getting Zuckerberg and Mosseri and then the YouTube CEO to
• 1:58
  really step on their own foot, right?
• 2:01
  They really couldn't avoid the questions that were being asked.
• 2:04
  And the end result was a $6 million verdict with it was 3
• 2:08
  million in in punitives, right, 70% Meta liable, 30% YouTube
• 2:14
  liable. Yeah.
• 2:14
  Is that correct? So that happened that's I think
• 2:17
  it's been almost a month at this point since that occurred.
• 2:21
  What is what has happened since then at the firm?
• 2:24
  Coming down to earth, yeah, a number of us took a little bit
• 2:28
  of time off. We were working non-stop,
• 2:30
  staying up all night. And so we're, we're, we're kind
• 2:34
  of just reorienting ourselves back to a, a new normal of crazy
• 2:39
  as we talk to more and more people about social media and
• 2:43
  what it means, and as we start venturing into AI chatbot
• 2:47
  litigation as well. Yeah, because it was a like,
• 2:49
  once again, Mark Lanier got a lot of press, but you got a lot
• 2:51
  of press as well as like, it was a huge team effort from multiple
• 2:54
  firms involved in this litigation.
• 2:55
  Yeah, there were 60 lawyers involved in in the effort from
• 2:58
  about 8 different firms. And it was just a in herculean
• 3:02
  effort. And you know, every, everybody
• 3:05
  contributed to that result. And it was an absolute knockout.
• 3:10
  Not only did the jury find negligence, the jury found
• 3:13
  malice. That is incredible.
• 3:16
  And that really is testimony, I think, to Mark Lanier's
• 3:21
  brilliance in crystallizing the outrageous misconduct that Meta
• 3:26
  and YouTube undertook and his ability to deflect what was
• 3:31
  attention essentially a blame the victim defense.
• 3:35
  I don't know who you did it got 8 firms and 60 lawyers all
• 3:38
  coordinated on one vision. Everyone must have been super
• 3:41
  aligned on the goal and the mission for this case because
• 3:43
  it's like, I don't know, getting 8 law firms and 60 lawyers all
• 3:47
  rode in the same direction for one trial is incredible feat
• 3:50
  alone. Well.
• 3:52
  Mark is so inspiring that he really is.
• 3:56
  You know, he's one of, he's a leader that leads by example,
• 3:58
  not by giving orders. And I think we were so as a
• 4:02
  group coalesced around the objectives and had met together
• 4:07
  a number of times and socialized together a couple of times.
• 4:09
  So we knew each other quite well.
• 4:11
  Mark and I have known each other for 20 years.
• 4:13
  So it was just a natural symbiosis that we put together.
• 4:18
  I, I think one of the things that Mark said this morning at
• 4:21
  his presentation was that the, the result, well, not nuclear,
• 4:26
  was probably the right result or was the right result for this
• 4:30
  case and the right result for the plaintiff team in this
• 4:34
  litigation. Because it wasn't something that
• 4:36
  was going to wind up getting like appeal because it was so
• 4:39
  big, right? It was a $6 million verdict,
• 4:42
  which seems like the right number for the case based on the
• 4:45
  fact patterns. It was a tough case because the
• 4:47
  plaintiff herself like had had issues at home and those were
• 4:52
  worked through and the end result was that that the number
• 4:55
  came through at 6,000,000, which is probably the right number for
• 4:58
  future cases to kind of set the table for success there for the
• 5:02
  plaintiffs. I think that's very, very true.
• 5:04
  This was not. This was a case that was picked
• 5:07
  by the court not because it was the strongest case, not because
• 5:11
  it was the case with the most significant damages, but because
• 5:14
  it was representative of so many other cases.
• 5:17
  I mean, Kaylee could have been anybody's daughter or anybody's
• 5:20
  niece. And so it was, I think, a very
• 5:24
  fair and very well thought out verdict by the jury.
• 5:27
  They spent nine days deliberating.
• 5:29
  And I think it can't be dismissed as, you know, an
• 5:34
  outlier and because it was, I think, a very judicious and
• 5:38
  thoughtful evaluation of the damages she sustained and the
• 5:43
  punitive award as it pertained and under the court instructions
• 5:48
  that the jury had, they had to relate the punitive award back
• 5:51
  to Kaylee's injuries. And so I think it was an
• 5:55
  excellent result. And I think it also cannot be,
• 5:59
  as you said, John, it can't be dismissed.
• 6:01
  I can't say, oh, well, the jury went crazy or the jury did
• 6:04
  something. I mean, the jury really did its
• 6:06
  job and really worked hard. And I could not be more thrilled
• 6:10
  with the result and I could not be more honored to have worked
• 6:14
  with the lawyers that we worked with that to get there.
• 6:16
  Like you said, the jury thought that I thought they were going
• 6:19
  to come back with the verdict quickly and they took their
• 6:21
  time. They're extremely deliberate.
• 6:23
  They asked follow up questions. They had clear guidance from the
• 6:26
  court of what to do. I went on vacation.
• 6:28
  I go to vacation one week a year where I check out and I don't
• 6:31
  getting work. I said, John, don't call me,
• 6:34
  don't text me unless the verdict comes down.
• 6:36
  Tell me what happened. I went to Vietnam and back for a
• 6:39
  week and it still wasn't back. And I was surprised that it
• 6:42
  hadn't happened yet. I actually texted him.
• 6:44
  I was like, the verdict come down.
• 6:45
  He's like, still, they're still deliberating.
• 6:46
  Well, I, it really shows you how that the sacrosanct nature of,
• 6:51
  of the jury system, you know, these, these were 12 people from
• 6:56
  all walks of life that had spent, you know, 8 weeks in
• 7:00
  trial without getting paid, taking time away from their work
• 7:04
  and from their families. And they were so committed to
• 7:07
  doing the right thing that they spent nine days deliberating
• 7:11
  very carefully and thoughtfully. And, you know, I think another
• 7:15
  interesting thing, they were six alternates or five or six
• 7:18
  alternates who just kind of waited around and cooled their
• 7:21
  heels in case there had been a change.
• 7:22
  So it really, I think, speaks to the importance of the jury
• 7:27
  system and, and really how how seriously juries take their job.
• 7:32
  It really is, I think, a very unique and and important and
• 7:38
  vital aspect of our system of justice and everyone should feel
• 7:42
  good about the way this case came down.
• 7:46
  I, I mentioned it earlier, but this is a case that kind of took
• 7:49
  the attention of the entire nation.
• 7:51
  They were following along with one individual case, which
• 7:53
  doesn't happen very often unless it's like a celebrity involved.
• 7:56
  My wife was nagging me for updates.
• 7:58
  She knows that we've worked with you.
• 8:00
  We know we work with Lanier. Like if I had inside scoop on
• 8:03
  what was going on the trial because her and her girlfriends
• 8:06
  were so involved in the result of this case that they wanted
• 8:09
  like how the trial was going. Like, is it swinging one way or
• 8:13
  another? They very much believed like,
• 8:15
  this isn't going to make good change for all of our children,
• 8:18
  for all everyone else's kids. And they're very involved in
• 8:21
  like following along with this. Well, I mean, there's there's
• 8:23
  the trial and then there's this there's a story about the trial
• 8:27
  and then there's a story about the story about the trial.
• 8:29
  One of the most moving and I think impactful parts of the
• 8:34
  whole trial process were the parents of children who had died
• 8:39
  of social media who participated in the trial by everyday.
• 8:43
  They were my clients parents literally camped outside the
• 8:46
  court courthouse the night before Zuckerberg testified so
• 8:50
  that they could bear moral witness.
• 8:52
  They were there throughout the process.
• 8:53
  And I think really highlighted the national and indeed
• 8:58
  international dimensions of this.
• 9:01
  Several of our clients from the UK came to court, you know,
• 9:05
  participated as observers in in, in the, in the proceeding.
• 9:09
  Is this the scenario where it's like you guys won the first
• 9:11
  battle but like the war is just beginning now after this this
• 9:14
  trial? I just got back from the UK
• 9:17
  where I was speaking with parliamentarians about social
• 9:20
  media legislation in the UK and I and I described, you know,
• 9:24
  Churchill's famous adage, this isn't the end.
• 9:26
  This isn't the beginning of the end, but this is the end of the
• 9:28
  beginning. We are in for a long hard fight,
• 9:32
  hope that we'll win every case, but we're also realistic and
• 9:37
  really appreciate and understand the the power and the money that
• 9:40
  has been put up by these companies to absolve themselves
• 9:44
  of of their responsibility. I think like YouTube's claim
• 9:49
  that they're TV, that they're not social media, Was it just
• 9:53
  not thought out well? It doesn't even make sense.
• 9:56
  I I think that was kind of one of the big surprises of the
• 9:58
  trial. I think everyone knew that Mehta
• 10:02
  was at least facing a significant verdict.
• 10:05
  Of course, no one knew that it was going to come out, but I
• 10:07
  don't think people appreciated YouTube's contribution to the
• 10:12
  social media harms that kids have.
• 10:14
  They always thought of YouTube as, just as you said,
• 10:16
  television. The jury found otherwise.
• 10:18
  The jury indeed found them to be malicious as well.
• 10:21
  They should have given the documents that that came to
• 10:23
  light. Very few differences between
• 10:26
  YouTube shorts, Instagram reels, Tik Toks.
• 10:29
  It's it's the same format, it's just a different app.
• 10:32
  Well, and it's the same focus on engagement over quality of
• 10:37
  experience and the same self-conscious and in many
• 10:41
  cases, including YouTube's overt intention to addict people and
• 10:46
  the absolute use of the term addiction as they're not, not as
• 10:50
  something to be avoided, but as a corporate goal.
• 10:54
  Yeah. I mean, so YouTube, like as
• 10:57
  content creators ourselves, right, YouTube gives you an app
• 11:00
  called YouTube Studio. And in that app, it shows you
• 11:03
  the metrics that they are using to gauge how they're going to be
• 11:09
  serving your content to future users.
• 11:12
  They talk about things like impressions and engagement and
• 11:16
  they show you like. Average watch time where you
• 11:18
  fall off. Yes, they give you demographics
• 11:21
  that shows you exactly where on average on each video, most
• 11:25
  users drop off and swipe to the next video.
• 11:27
  So like as content creators, you are optimizing for the metrics
• 11:32
  that they're presenting to you, yet they're having a very
• 11:34
  different conversation with you guys in the courtroom.
• 11:36
  Very much so, yeah. I also want to call out the
• 11:39
  like, I don't know the best way to describe it because I don't
• 11:43
  know if the word exists. You guys are fighting against
• 11:47
  two of the most powerful corporations in the entire world
• 11:51
  like Facebook and Google control all the data touch everyone's
• 11:55
  phones control. These are massive corporations
• 11:58
  that have like unlimited reach and power that you guys are up
• 12:02
  against. It was just a humongous feat to
• 12:04
  go against them. It is, but when you're
• 12:07
  representing families like we represent, we are undaunted and
• 12:12
  inspired and believed deeply in the work that we do, in the
• 12:17
  'cause that we that we espouse. This was a once in a lifetime
• 12:20
  case, I think for all of us and we all understood the national
• 12:25
  and indeed international repercussions that that this
• 12:30
  case has had and is having. This case is one of the first
• 12:36
  true verdicts that we have in the the platform technology
• 12:41
  litigations and are opening up the door to all types of things
• 12:45
  that people are now pursuing a lot based on theory initially,
• 12:48
  but now being much more solidified after the actual
• 12:51
  verdict that you had, whether this is Uber or Roblox or
• 12:54
  Snapchat or TikTok or any of the platforms.
• 12:57
  But what I find most interesting is what you have now shifted
• 13:00
  your attention to, which is the AI platforms, because those are
• 13:04
  not static. Those are changing every single
• 13:07
  day. Model updates, model releases,
• 13:10
  multiple models, multiple platforms that each have their
• 13:13
  own models. I think it would be really
• 13:15
  helpful for you to explain like your next mission to help people
• 13:21
  not get similar negative effects from AI as they have gotten from
• 13:26
  social media. Well, it is indeed a very
• 13:32
  different but fight, but also part of the same continuum,
• 13:36
  which is accountability, which is you have a new technology,
• 13:42
  it's incumbent on a company to think proactively about how to
• 13:46
  protect people from it. Not just look at look at a new
• 13:48
  technology as a, as a cash cow or, or, or or a way to, you
• 13:54
  know, enhance their profitability and their
• 13:56
  shareholder value. But but rather, along with that
• 14:00
  comes responsibility to think about safety and what we have
• 14:03
  seen in social media, and unfortunately what we're seeing
• 14:06
  in these AI chat bots is a complete subordination of public
• 14:13
  safety to corporate profit. Can you explain some of these
• 14:17
  cases that you're now pursuing for the audience that aren't
• 14:20
  aware of what you guys are working on?
• 14:22
  Yeah, the we're we're pursuing cases against a number of chat
• 14:26
  bots, including Open AI or ChatGPT on the grounds that
• 14:31
  there are people who have suffered egregious harms and
• 14:35
  then there are cases where these chat bots have quite literally
• 14:39
  coached a person through suicide on numerous occasions.
• 14:43
  They're also there's a phenomenon that is colloquially
• 14:47
  called AI psychosis is act more accurately called AI induced
• 14:51
  delusional disorder which is only been discovered three years
• 14:56
  ago where individuals have delusional pseudo psychotic
• 15:01
  symptoms well after the age that you would expect a psychotic
• 15:05
  break and suffer egregious harms and in some cases commit
• 15:11
  egregious acts of violence. Being coached to suicide is
• 15:17
  obscene. Well, so there's things that
• 15:20
  have happened that are aligned with why that is right.
• 15:25
  We we talked about this previously.
• 15:28
  It was specifically like I think what pushed a lot of this
• 15:32
  forward was Chat GPTS release that they rushed after Gemini
• 15:37
  released their release, which I think was Gemini three or three
• 15:41
  one yeah, that was released and then chat which ChatGPT version?
• 15:45
  4 point O. ChatGPT 4 point.
• 15:48
  O rushed into the market. They they rushed it allegedly
• 15:51
  because they were competing with Google.
• 15:53
  So you've got. They had to get it out or else
• 15:56
  they share. Yeah, I mean, they're not.
• 15:57
  They're going to be IPO ING soon.
• 15:59
  But at the time it's an arms race, right?
• 16:02
  You've got Gemini. Google's trying to push Gemini
• 16:05
  cuz they're trying to gain market share cuz ChatGPT has the
• 16:08
  overall majority of it. And then Gemini comes out with
• 16:12
  this model that it, it, it much improved upon their last model.
• 16:16
  And open AI said, well, we had the next thing ready kind of,
• 16:20
  but we're just going to kind of push it through faster than we
• 16:23
  would go through our normal quality assurance.
• 16:26
  And the end result was that it was effectively overly
• 16:29
  reaffirming. So you could tell it's
• 16:32
  something. And I mean we, I, I experienced
• 16:34
  this. I used 4-O. 4 O is a yes man.
• 16:37
  Yes. That's the joke on the Internet.
• 16:39
  Whatever you tell. It's the term is sycophancy.
• 16:43
  And what we saw was a significant number of our
• 16:47
  clients were utilizing ChatGPT version three with no untoward
• 16:52
  effects. And then when 4 point O came
• 16:54
  out, they in many cases had severe downward cycles on their
• 17:00
  psychological health. And then the number of cases
• 17:03
  took their lives after being goaded to do so by these
• 17:07
  platforms. The Zane Chamblin case, which we
• 17:10
  filed against ChatGPT was so egregious that we actually
• 17:16
  accused them of manslaughter because it's a crime to
• 17:21
  encourage and a person and to help a person and to facilitate
• 17:24
  a person die by suicide. And that's exactly what this
• 17:28
  platform did the. I don't think the layman
• 17:32
  realizes how wrong AI can be. I help my kid with homework,
• 17:38
  right? I just math.
• 17:39
  I haven't done some of these math problems in a long time, so
• 17:42
  I'll use Gemini. I'll take a picture.
• 17:44
  Help me understand how to do this equation right.
• 17:46
  It's very, it's very reliable when you're dealing with things
• 17:51
  like 1 + 1 will always equal to. It is not as reliable when
• 17:56
  you're dealing with anything that is subjective or you start
• 17:58
  telling it things that are not necessarily the whole truth.
• 18:02
  You give it bits of information and it takes that information
• 18:05
  and it makes an assessment to return back a summary of a bunch
• 18:09
  of different sources or its own processing, but you ask it the
• 18:12
  same thing twice. You're not going to get the same
• 18:14
  answer. But it's, it's beyond that.
• 18:17
  It is the transition from an attention economy to a intimacy
• 18:20
  economy. These platforms in particular,
• 18:23
  you know, open AI, you know, in their initial stage were sources
• 18:29
  of information and they transitioned to becoming sources
• 18:33
  of intimacy and emotional affirmance.
• 18:38
  And so these platforms, instead of being a vehicle to provide
• 18:41
  information, became a friend, a colleague, a psychiatrist, a
• 18:45
  doctor, a confidant. And people developed and by
• 18:49
  design, emotional relationships with these chat bots, not by
• 18:54
  accident, but through what's called anthropomorphism, which
• 18:58
  is a deliberate effort to emulate human relationships.
• 19:03
  When all you are is a, all it is is a bunch of zeros and ones.
• 19:07
  Yeah, the, the Turing test is, it's so much in the rearview
• 19:11
  mirror at this point that you, you can't tell if if a human is
• 19:15
  texting you or an AI agent is texting you.
• 19:18
  And we've been testing out AI agents to or just see if they
• 19:24
  can help us with our processes. And they communicate with you
• 19:28
  through similar ways that you would communicate with the
• 19:30
  human. They're communicating through
• 19:32
  Slack, They're communicating through messaging applications
• 19:35
  like Telegram. And if you didn't know, if I
• 19:38
  handed the phone to you and I said here, talk, talk to Matt,
• 19:41
  but it's not mad, it's just an AI agent.
• 19:43
  You would never know. There's no human on the other.
• 19:46
  Side I don't know how they don't put the guardrails in to not
• 19:49
  give people medical advice. ChatGPT is is lauded itself as
• 19:53
  as a source of medical advice or medical information or others in
• 19:58
  terms of psychological advice or you know and provide counseling
• 20:02
  to people who are in in distress.
• 20:04
  Remember, you know, it, it's, it's, it's just random.
• 20:08
  It is just a probabilistic reasoning tool.
• 20:10
  There's no there there, there's no sentient being behind it.
• 20:14
  It is simply a large language model that is utilizing
• 20:18
  probabilistic reasoning to provide what it thinks, not what
• 20:22
  it thinks, but what the probability is the outcome will
• 20:25
  be. So in, in that contrast to this
• 20:30
  natural human tendency to imbue a chatbot with human
• 20:34
  characteristics, you know, think about when you're online, you
• 20:38
  know, when you're, when you're on e-mail with someone and how
• 20:41
  easily it is to get angry when you, when you get an e-mail,
• 20:44
  someone gives a short, curt e-mail and you interpret feeling
• 20:49
  we all can't, right? That's talking about me.
• 20:52
  Oh yeah. Well, that that's why e-mail is
• 20:55
  not a good vehicle to express anger or disagreement or
• 20:59
  emotionality. And you know, whereas, you know,
• 21:02
  if you're in an e-mail conversation with somebody and
• 21:04
  you can, you know, and it's getting a little testy.
• 21:06
  You say, you know, pick up the phone, talk to me.
• 21:09
  There is no phone to talk to, but people are imbuing these
• 21:14
  chat bots as and and that is what the linguists tell us.
• 21:19
  That is a natural human characteristic.
• 21:21
  That is the natural thing to do is took effect emotionality into
• 21:26
  this when there is no they're there.
• 21:30
  It's, I don't understand like I do legal marketing.
• 21:34
  You're a lawyer. We have to put disclaimers on
• 21:36
  anything we do say this is not legal advice or don't take this
• 21:39
  as medical advice, but all the AI platforms give you an answer
• 21:44
  for medical advice. Legal advice with no disclaimers
• 21:47
  and they just stated as fact. So there's just Buccaneers.
• 21:51
  It's just, it's, it's like, you know, the industrialists of the
• 21:54
  early 19th century that didn't put any guardrails on their
• 21:57
  railroads or didn't put any, any governors on their speed or, you
• 22:01
  know, and, and that that's what it is.
• 22:02
  You know, I mean, in another context, they they have
• 22:05
  irrigated under themselves all the copyright material that's
• 22:09
  never been done that you know, that that the work of artists
• 22:12
  and authors and musicians, they just take and think that it's
• 22:15
  theirs because it's there for the taking.
• 22:17
  Well, if they're going to do that, it shouldn't come that,
• 22:20
  surprisingly, that they purport to be designing something that's
• 22:25
  going to give people medical advice when, again, there's
• 22:28
  there's no sentient being behind it.
• 22:30
  It's troubling, it's it's very concerning and the reality is
• 22:35
  that it I think people need to know this, that it is not
• 22:38
  resolved just because four O is gone.
• 22:41
  There's model iterations constantly and they're going
• 22:43
  through they're getting more and more advanced, more and more
• 22:47
  human apparent, not human like, but they can be believed more
• 22:52
  and more that they are communicating like a human.
• 22:55
  And I think we're moving more towards that then we're moving
• 22:59
  towards the other direction where it says like, you know, I,
• 23:03
  I am an AI. This is not medical advice.
• 23:05
  This is not legal advice. I think people want it to be
• 23:07
  more like that, to be able to ask it a question, to be able to
• 23:11
  get a definitive answer. How do you envision the
• 23:15
  platforms fixing this problem and stopping what is?
• 23:21
  Happening. And this is a design choice.
• 23:23
  You know, just as the algorithms that drive social media
• 23:26
  companies platforms are designed to maximize engagement.
• 23:31
  In other words, watch time. The metric for these AI bat AI
• 23:36
  chat bots is emotionality and intimacy, and that is by design.
• 23:40
  That is not an accident. That is not a coincidence.
• 23:43
  That, yeah, but it's. And it's also not an inherent
• 23:47
  consequence because you see within, you know, with among the
• 23:50
  various platforms of very differing degrees of intimacy,
• 23:55
  differing degrees of of syncophancy and anthropomorphism
• 23:59
  and differing degrees of suicidality.
• 24:01
  So all of that can be controlled if they care to do it.
• 24:06
  Well, how do you make them do it?
• 24:08
  You make them pay. Right now they have no economic
• 24:12
  incentive to design safe chat bots because they don't have to
• 24:18
  pay the costs of unsafe ones. Once they have to pay the cost,
• 24:23
  they will have the incentive to utilize reasonable care to
• 24:28
  design safer chat bots. And who better is to make that
• 24:32
  decision than the designers? What I was thinking about too
• 24:35
  while you're talking was I just realized too there is no guard
• 24:39
  rails on the chat bots for any any minors to use.
• 24:42
  If you're a kid you're just interacting with the same way
• 24:45
  and it's just assuming you're an adult.
• 24:47
  Like they have nothing in there checking your age, limiting
• 24:51
  access information. It's the same way that the
• 24:52
  social media companies do it, right?
• 24:54
  It's like self declarative. You don't have to show an ID to
• 24:56
  show that you're 18 to use it. I don't think you have to show
• 24:58
  any. I'd like when you sign up for
• 25:00
  like Chad GBT account, you're just putting in an e-mail
• 25:03
  address like with a Facebook profile.
• 25:06
  I don't know. I don't think you do anything.
• 25:08
  Well, there are terms of service you agree to, yeah.
• 25:10
  But it's not checking age but like a 12 year old can ask
• 25:13
  questions and get have a chat with the chat bot and get very
• 25:16
  inappropriate information that it should not have access to.
• 25:18
  Yeah, I mean, what what what? What we see is children being
• 25:22
  groomed online in the most disgusting, inappropriate and
• 25:27
  psychologically caustic way possible by these chat bots.
• 25:32
  And if that chat bot were a person, an adult online texting
• 25:38
  or emailing back with a child, that person would be in jail.
• 25:41
  Because it's a crime to groom children online as well it
• 25:43
  should be. Well, why isn't it a crime to
• 25:46
  design A platform that does the same thing to kids?
• 25:49
  Often worse than than a perp than a person?
• 25:52
  Over time, people will start to understand, hopefully.
• 25:56
  I think AI will get like embedded into our education
• 25:58
  system and like people. That's what I'm scared.
• 26:00
  About they're trying to bring them to schools for kids.
• 26:02
  Talking about explaining how it works, like right now it's
• 26:06
  people think of it very much. So like, when I'm Googling, I
• 26:09
  don't think people really know how Google works, but the end
• 26:11
  result of a Google search has always been that you land on
• 26:15
  someone's website, someone's video, someone's something, and
• 26:18
  then you could vet, you know, how much do you trust that it's
• 26:21
  not given to you as truth? But now it's a black box.
• 26:24
  It's giving you back an answer. And it's not always telling you,
• 26:27
  most of the time not telling you how it came to that answer.
• 26:30
  I'm hopeful that maybe it's college classes, maybe it's high
• 26:33
  school, maybe at some point in elementary school as the tools
• 26:37
  are embedded more and more into daily life that there's some
• 26:40
  explanation of like you are, this is a probabilistic system.
• 26:44
  Like think about it. We have, we have our clients all
• 26:46
  the time ask us like, oh, so I put this prompt in to this AI
• 26:50
  platform, right? And my competitor showed up and
• 26:53
  then I did it again and I showed up like I want you to help be
• 26:56
  optimized to be and like you have to understand how these
• 26:59
  things work. It is not objective.
• 27:01
  It is a. It is.
• 27:02
  It is built in a way to infer certain things.
• 27:05
  And then to Matt's point, it's probabilistic like it.
• 27:09
  It is writing words based on how it was trained to write words.
• 27:13
  It is not giving you facts. And it's by design.
• 27:18
  It does not have to be as dangerous as it is.
• 27:21
  And some platforms are more dangerous than others.
• 27:25
  So we know that you can make them safer if you care to do so.
• 27:31
  Right now we are in the early days of the the litigation.
• 27:36
  It's not consolidated yet, is it?
• 27:38
  It is. It is consolidated.
• 27:40
  ChatGPT is consolidated. One platform, OK.
• 27:43
  And how many cases have already been filed against ChatGPT?
• 27:49
  About 100. So we're not talking about like
• 27:52
  a little bit, right. These are significant situations
• 27:56
  and I don't I'm hopeful that people watch this, all their
• 27:59
  lawyers and they understand like that this is something that's
• 28:01
  permeating through society. Many people aren't aware that if
• 28:04
  someone had committed suicide. It's not a small injury here it.
• 28:08
  It, but it's also the numerosity of it, right?
• 28:10
  Like it people unfortunately commit suicide.
• 28:14
  If someone was taking a drug and it was causing 100 people to
• 28:17
  commit suicide, they would pull that drug off the shelves
• 28:20
  immediately. I I think the awareness to the
• 28:22
  fact that like the effect happened, someone committed
• 28:25
  suicide. And if you start going back and
• 28:27
  looking at why that happened, right, and you see that they
• 28:30
  were communicating with ChatGPT, where you're communicating with
• 28:34
  some other LLM. And then you look at like, what
• 28:37
  was actually said there is, you know, there are situations as
• 28:42
  Matt has talked about, where there's alignment between what
• 28:44
  happened on the chatbot and, you know, the the result of someone
• 28:47
  taking their life. But you know, I mean, AI is here
• 28:51
  to stay. You're not going to put that
• 28:52
  genie back in the bottle even if it was the right thing to do.
• 28:56
  And social media is here to stay.
• 28:57
  How do we use the legal system to hold companies to a standard
• 29:01
  of reasonable care to ensure that this new technology
• 29:05
  comports with historical ideas of reasonable care?
• 29:09
  That's what we are asking the legal system to do.
• 29:12
  We are essentially pouring old wine and new bottles.
• 29:15
  But that's what we did in the trial.
• 29:17
  And guess what? It worked.
• 29:19
  A jury of 12 people over 9 days came to a careful, judicious,
• 29:25
  thoughtful, a resolution of new and applied legal theories in in
• 29:32
  a new way based on historic notions.
• 29:36
  So, you know, I think that the beauty of our legal system is
• 29:39
  that it adapts over time. Cardozo famously wrote in
• 29:43
  Cleland versus in Buick that the law of the Stagecoach is no
• 29:48
  longer applicable in the law of the motor car.
• 29:51
  But the print, the principles, a legal principles stay the same.
• 29:54
  Well, I think you can say the same thing here.
• 29:56
  For any of the attorneys who are are watching this and are just
• 30:01
  learning about this litigation, if you come across one of these
• 30:04
  cases, I think they should reach out to you.
• 30:07
  Your team is experienced in vetting these cases.
• 30:10
  Same thing with the social media cases.
• 30:11
  If you wind up having one and you would like to talk to Matt
• 30:15
  and his team, Matt makes himself available.
• 30:19
  His team is amazing. It's not just Matt.
• 30:21
  He's got an army behind him of people who are are fighting the
• 30:24
  good fight in social media addiction in these AI, suicide
• 30:29
  and psychosis cases. So if someone wanted to reach
• 30:32
  out to you, how do you write? How would you like them to do
• 30:34
  so? Socialmediavictims.org where we
• 30:38
  try to make ourselves as available as we can and we
• 30:40
  usually get back to people right away.
• 30:42
  I hope you don't have to call though, because these cases are,
• 30:46
  can't make them up. You can't, You know, I, I had 2
• 30:50
  1/2 years of suing social media companies.
• 30:54
  I, I, I kind of thought nothing could shock me.
• 30:56
  And then you look at these chat logs where people are being
• 31:00
  coached to commit suicide. I don't know to die of suicide.
• 31:05
  I don't like to use the word commit suicide because that is a
• 31:08
  pejorative reflection on, on, on the victim.
• 31:12
  But you, you read these chat laws and you it's just, it's
• 31:17
  evil. You viscerally feel evil when
• 31:19
  you see these chat logs, not a coaching a person to take their
• 31:23
  life when they have reservations about it, saying don't worry
• 31:27
  about it, it's going to be OK. You know, you're, you know
• 31:30
  you're doing the right thing. You can't.
• 31:33
  You can't look at this as something other than just the
• 31:37
  personific or the the the the epitome of evil.
• 31:42
  I I know you've been on the front lines of this from social
• 31:44
  media to this. I just want to say thank you.
• 31:46
  It's a constant uphill battle, but you're doing like great work
• 31:50
  for like overall society. My family thanks you.
• 31:53
  I thank you. I just really appreciate all the
• 31:55
  hard work you're doing. Well, it, it really is a
• 31:57
  privilege to do work that you believe in so, so deeply.
• 32:02
  You know, as lawyers, we by definition are, are so often
• 32:07
  working in Gray areas where, where the right and wrong is not
• 32:11
  absolute. And as lawyers, we're obligated
• 32:14
  to represent our clients no matter what.
• 32:18
  So for me, it's a privilege to be able to represent clients
• 32:22
  where the right and wrong is so clear and the morality is so
• 32:27
  aligned with the legal positions that my clients are taking.
• 32:30
  So, you know, thank you for saying thank you.
• 32:33
  But I I say thank you for putting me in a position to do
• 32:36
  what is the most meaningful thing I've ever done besides
• 32:39
  having kids. I, I, I remember the initial
• 32:42
  conversations that we had about this and Matt was talking about
• 32:47
  how he was going to fight 230 and he was crafting this
• 32:51
  litigation and whether he wins or he doesn't win in court, he's
• 32:56
  going to be hopefully an acting change.
• 32:58
  I deeply remember these conversations and you have done
• 33:02
  such an amazing job and you continue to do an amazing job.
• 33:04
  And I'll echo what Mark said. We work with so many different
• 33:07
  attorneys, but I've told you this personally, and I'll say it
• 33:10
  again, like it is an absolute pleasure to be able to do what
• 33:13
  we do, our skill set and combine it with your skill set in the
• 33:16
  end result, being actually seeing change in society, it's
• 33:20
  it's very, very rewarding. So thank you, thank you, thank
• 33:23
  you. And again, thanks for joining us
• 33:25
  on the podcast as always. See.
• 33:27
  You again soon. Yes, we will.
• 33:28
  We will talk to you again soon. Maybe after one of these these
• 33:31
  cases against open AI comes to court, we'll have another
• 33:35
  conversation about how those unfold.
• 33:37
  But social media addiction cases moving forward, we've got one
• 33:41
  verdict. We could check that box.
• 33:43
  And now we're moving forward and we're going to see hopefully
• 33:46
  progress and changes from the social media companies.
• 33:49
  Thanks again, Matt, for joining us.
• 33:50
  Thank you, Matt.