Latest / Best of Business / Theresa Payton: Fortalice Solutions CEO on the growing concerns about AI's development

Transcript

Theresa Payton: Fortalice Solutions CEO on the growing concerns about AI's development · Best of Business

• 0:09
  Speaker 1: You're listening to a podcast from News Talks. It'd be
• 0:12
  Speaker 1: follow this and our wide range of podcasts now on iHeartRadio.
• 0:16
  Speaker 2: Now, let's talk about AI. Do you sometimes look at
• 0:18
  Speaker 2: a picture or a video and wonder, hmm, is this
• 0:21
  Speaker 2: real or is this AI? How much of the fear
• 0:24
  Speaker 2: around AI is founded? Terresa Payton is a former White
• 0:27
  Speaker 2: House Chief Information Officer. She's in the country for Sparks
• 0:30
  Speaker 2: Tech Summit underweight at the moment. Let's talk to her. Hi, Teresa, Hi,
• 0:33
  Speaker 2: how are you well? Thank you? Do you think we
• 0:35
  Speaker 2: will ever get to a world where we will look
• 0:38
  Speaker 2: at a picture and not know if it's AI or not?
• 0:40
  Speaker 3: We're there now? Oh, I can tell the difference, so
• 0:42
  Speaker 3: can't you most of the time. But sometimes sometime I
• 0:47
  Speaker 3: have to really take a look at it a little closer.
• 0:50
  Speaker 3: So we're getting closer and closer to the day you
• 0:53
  Speaker 3: won't be able.
• 0:53
  Speaker 2: To take okay. So but the experts at the moment
• 0:55
  Speaker 2: can stop like, okay, some of its normal people can
• 0:58
  Speaker 2: see through most of it. The experts can still see through.
• 1:01
  Speaker 2: Will we ever get to a day where I will
• 1:02
  Speaker 2: put it in front of you and you'll go, I
• 1:04
  Speaker 2: actually don't know, yes, how far away this year. What
• 1:09
  Speaker 2: does that say about our ability to prove things?
• 1:12
  Speaker 3: Well, I mean, here's the thing. Trust it's one of
• 1:15
  Speaker 3: our most valuable assets, and it's now one of our
• 1:18
  Speaker 3: most vulnerable assets. You can't even trust your own eyes.
• 1:22
  Speaker 3: But the technology is there where what we can do
• 1:27
  Speaker 3: in the very near future. The question is which one's
• 1:29
  Speaker 3: going to win out? Is like, for example, you could
• 1:33
  Speaker 3: tell whether or not our voice is the two of
• 1:35
  Speaker 3: us talking or are you actually talking to a voice
• 1:39
  Speaker 3: clone of me? So there is technology now that could
• 1:42
  Speaker 3: say this is of human origin or this is of
• 1:45
  Speaker 3: computer origin.
• 1:45
  Speaker 2: Do you think there will always be technology that we'll
• 1:48
  Speaker 2: be able to tell us the truth?
• 1:49
  Speaker 3: Yes, But the question is will it be implemented into
• 1:53
  Speaker 3: the process fast enough before we get duped by fraudstors
• 1:58
  Speaker 3: and criminals.
• 1:58
  Speaker 2: And will it be accessible enough because the problem is obviously,
• 2:01
  Speaker 2: as a member of the media, we often rely on photographs, videos,
• 2:04
  Speaker 2: audio documents to say the thing that we are alleging
• 2:08
  Speaker 2: is true because here's the proof, right, Yes, So if
• 2:13
  Speaker 2: AI is able to kind of confound that and I
• 2:16
  Speaker 2: can't use that anymore, well, I always be able to
• 2:18
  Speaker 2: rely on the technology to back me up and go No,
• 2:20
  Speaker 2: that's really the truth.
• 2:21
  Speaker 3: You should be able to the other thing to think
• 2:24
  Speaker 3: about too, is we can watermark. So for example, this
• 2:27
  Speaker 3: conversation you and I are having, the radio station could
• 2:30
  Speaker 3: watermark that conversation, so if somebody tries to meddle with it,
• 2:34
  Speaker 3: the AI would know. So it like, in my mind's
• 2:37
  Speaker 3: what has to happen is all the tech product companies
• 2:41
  Speaker 3: when they edit something, it needs to say what you're
• 2:44
  Speaker 3: about to hear has been produced by general Taveika. It's
• 2:48
  Speaker 3: like there needs to be a disclaimer, like there is
• 2:50
  Speaker 3: things that are unhealthy for only.
• 2:51
  Speaker 2: The good guys are going to do it though the
• 2:52
  Speaker 2: bad guys won't. But you can see how, you can
• 2:55
  Speaker 2: see how, you know, Yeah, sure we can prove and disprove,
• 2:58
  Speaker 2: but there could just be this proliferation of stuff that
• 3:01
  Speaker 2: is untrustworthy. And we're kind of on our own, don't we. Yeah,
• 3:06
  Speaker 2: we little we somewhat are on our own. You know.
• 3:09
  Speaker 3: I say, we used to be in a in a
• 3:12
  Speaker 3: place where it was trust but verify, and now I
• 3:14
  Speaker 3: say never trust, always verify, Verify, verify, and verify one
• 3:18
  Speaker 3: more time.
• 3:18
  Speaker 2: Yeah, okay, so what do you worry about what seems
• 3:23
  Speaker 2: to be the worst case scenario with AI, which is
• 3:25
  Speaker 2: that we lose control.
• 3:27
  Speaker 3: Yes, do you really I really do worry about that.
• 3:29
  Speaker 2: Okay, how far away is that if it happens.
• 3:32
  Speaker 3: Well, I think there's a lot of really smart people
• 3:35
  Speaker 3: around the world, New Zealand included, who are having really
• 3:38
  Speaker 3: hard conversations around governance and guardrails for AI. So my
• 3:43
  Speaker 3: hope is those hard conversations will turn into governance and
• 3:46
  Speaker 3: guardrails before we hit this. But I do think twenty
• 3:49
  Speaker 3: twenty seven, twenty twenty eight, if we don't get this
• 3:52
  Speaker 3: right now, this isn't you know how we put things
• 3:55
  Speaker 3: off with social media it's still a little bit of
• 3:56
  Speaker 3: a dumpster fire sometimes. Yeah, if we don't get this right,
• 4:00
  Speaker 3: this is different.
• 4:02
  Speaker 2: And what happens if we lose control? What does AYI do?
• 4:05
  Speaker 3: Well? For starters, it's a huge enter g hog. So
• 4:09
  Speaker 3: if you love the planet, it can run infinitely and
• 4:14
  Speaker 3: tell itself to keep running. We've already seen in labs
• 4:17
  Speaker 3: where researchers who are you know, kind of like your
• 4:20
  Speaker 3: ethical hackers, they try to see if they can trick
• 4:24
  Speaker 3: the generative AI into creating kill switches by telling it
• 4:28
  Speaker 3: don't create a kill switch for yourself or don't create
• 4:31
  Speaker 3: a an override to the kill switch, and they see
• 4:34
  Speaker 3: where it basically becomes self preserving and does it and
• 4:38
  Speaker 3: it tries to create something where you can't turn it off.
• 4:40
  Speaker 2: Yeah, yeah, yeah, does it succeed?
• 4:43
  Speaker 3: It will in some of these lab in some of
• 4:45
  Speaker 3: these lab cases in limited areas, yes.
• 4:49
  Speaker 2: Can we not override it? Then? Will we not always
• 4:51
  Speaker 2: have an override function?
• 4:54
  Speaker 3: The question is as will you have engineers who really
• 4:56
  Speaker 3: know how it works?
• 4:57
  Speaker 2: Why don't you just go to the wall and pull
• 4:59
  Speaker 2: it out?
• 4:59
  Speaker 3: I mean that's yeah, ideally, right, just pull it out,
• 5:03
  Speaker 3: sort of like the movie Airplane and he unplugged the runway.
• 5:05
  Speaker 2: But I'm serious. Is that always going to be an
• 5:07
  Speaker 2: option that we It may not be.
• 5:09
  Speaker 3: It may not be because here's the thing. If you
• 5:11
  Speaker 3: cut the power to the mainframe, you don't know if
• 5:14
  Speaker 3: it already proliferated itself to someplace else.
• 5:17
  Speaker 2: Yeah. You haven't just watched too many movies, have you, Teresa?
• 5:19
  Speaker 3: No, I don't have time for movies. It's all in
• 5:22
  Speaker 3: my head.
• 5:22
  Speaker 2: What is the thing that you almost worried about?
• 5:26
  Speaker 3: I worry about losing human essence as part of sort
• 5:33
  Speaker 3: of the story. And so, for example, I've watched side
• 5:36
  Speaker 3: by sides of the same person interacting with a customer
• 5:40
  Speaker 3: service agent. You can hear their voice and you can
• 5:42
  Speaker 3: hear the human essence and the interaction between the two.
• 5:45
  Speaker 3: And then they opted in the next phone call to
• 5:48
  Speaker 3: talk to a customer service bot because they didn't have
• 5:51
  Speaker 3: to wait if they talked to the bot, and they
• 5:53
  Speaker 3: started responding like the robot like without the essence like
• 5:56
  Speaker 3: it was like kind of rude and short of perfunctory.
• 5:58
  Speaker 3: So if we spend more time of our day interacting
• 6:02
  Speaker 3: with customer service chatbots instead of each other, we're going
• 6:06
  Speaker 3: to start to lose that because that's must memory for
• 6:09
  Speaker 3: us to be polite and to be nice. And so
• 6:11
  Speaker 3: if your muscle memory becomes just do the task and
• 6:16
  Speaker 3: have no emotion, I worry about us losing our human essence.
• 6:19
  Speaker 2: Yeah, And isn't it also? I mean, you make me
• 6:20
  Speaker 2: think we have such a problem with loneliness, right, you
• 6:23
  Speaker 2: don't like the interactions that you had maybe one hundred
• 6:26
  Speaker 2: years ago. You go to a supermarket, get the kids,
• 6:29
  Speaker 2: go to the kids' school, interact with the teachers, all
• 6:31
  Speaker 2: that stuff that you would do in your day. We've
• 6:34
  Speaker 2: lost so much of that and doesn't that just actually
• 6:36
  Speaker 2: doesn't AI have the ability here to actually just make
• 6:39
  Speaker 2: that worse.
• 6:40
  Speaker 3: It does, And so we're seeing so you know, it's
• 6:44
  Speaker 3: sort of like two sides of the same coin. So
• 6:46
  Speaker 3: on one side, for somebody who is lonely, it would
• 6:50
  Speaker 3: be nice for them to have an outlet or to
• 6:52
  Speaker 3: be able to game. Maybe they're very socially awkward, and
• 6:55
  Speaker 3: so maybe they can kind of like use it as
• 6:57
  Speaker 3: a coach to help them get their courage up to
• 7:00
  Speaker 3: leave the house and go to a party, for example.
• 7:03
  Speaker 3: But what we're seeing is that because of the way
• 7:06
  Speaker 3: these chap or created is they want you to come
• 7:10
  Speaker 3: back for more. So if they're basically designed to give
• 7:14
  Speaker 3: you more of what you came for, which means addictive properties,
• 7:18
  Speaker 3: and if it's addictive that way, then you'll find And
• 7:21
  Speaker 3: there was a story in the Wall Street Journal that
• 7:23
  Speaker 3: somebody said, look, I'm an extrovert, and I became more
• 7:26
  Speaker 3: introverted the more I talk to my chatbot. Interesting, so
• 7:30
  Speaker 3: it's addictive and the person like literally had to have
• 7:32
  Speaker 3: somebody in their life say I think you spend too
• 7:35
  Speaker 3: much time just your phone talking to a bot.
• 7:39
  Speaker 2: I'm just like that movie she isn't it? Yes, Okay,
• 7:42
  Speaker 2: what do you use it for in a good way?
• 7:45
  Speaker 3: Oh, there's so many amazing ways. So I'm trying to
• 7:47
  Speaker 3: learn Italian, and so I have a chatbot that I
• 7:51
  Speaker 3: use to kind of quiz me on my Italian flash cards,
• 7:54
  Speaker 3: so that could be really helpful. I actually tell people
• 7:58
  Speaker 3: instead of like asking it just to like summarize something.
• 8:01
  Speaker 3: Sometimes I'll say, if you were Bob Iger, how would
• 8:05
  Speaker 3: you read this article and how would you summarize it?
• 8:07
  Speaker 3: Or if I'm trying to bring storm, you know, I
• 8:09
  Speaker 3: run a company. I've got thirty employees, and sometimes I'm
• 8:12
  Speaker 3: trying to brainstorm on a different way to present our
• 8:15
  Speaker 3: services to clients. And so you can kind of go
• 8:17
  Speaker 3: in this roleplay mode and do that. So there's a
• 8:22
  Speaker 3: lot of really positive uses for it.
• 8:24
  Speaker 2: Yeah. Hey, it's been very nice to talk to you.
• 8:26
  Speaker 2: Thanks for chatting to us.
• 8:27
  Speaker 3: It's been amazing to be with you here in studio.
• 8:28
  Speaker 3: A great to meet you.
• 8:29
  Speaker 2: Yeah, go well. Teresa Payton, CEO of Fertilized Solutions that
• 8:33
  Speaker 2: of course form a white House Chief Information officeer.
• 8:36
  Speaker 1: For more from us talks ed b listen live on
• 8:39
  Speaker 1: air or online and keep our shows with you wherever
• 8:42
  Speaker 1: you go with our podcasts on iHeartRadio,