WEBVTT
00:00:05.921 --> 00:00:08.826
Welcome to the SMU Perspectives Podcast.
00:00:09.307 --> 00:00:12.089
I am Robert Ehlert, the SMU Commentary Editor.
00:00:12.109 --> 00:00:21.603
Our guest today is Robert Hunt, Director of Global Theological Education at the SMU Perkins School of Theology.
00:00:21.663 --> 00:00:32.398
Professor Hunt is the author of several books, but his latest delves into the intersection of artificial intelligence and the human species.
00:00:32.417 --> 00:00:35.502
It's called All Brain and No Soul.
00:00:36.130 --> 00:00:38.954
Real Humanity in an AI Age.
00:00:39.935 --> 00:00:50.572
Before we dive into that subject, Professor Hunt, elaborate how someone with a theology background began probing AI and its impact on our society.
00:00:51.292 --> 00:00:52.213
Thank you very much, Bob.
00:00:52.575 --> 00:00:53.155
A couple of things.
00:00:53.256 --> 00:00:57.802
First, I did actually have a long background in computers and computer science.
00:00:57.923 --> 00:01:00.225
I started as a computer programmer in university.
00:01:00.646 --> 00:01:02.289
Then I switched to history and theology.
00:01:02.722 --> 00:01:06.126
And so I followed this very closely over the course of my career.
00:01:06.947 --> 00:01:13.174
What brought this more immediately to mind was a conference we had at SMU about six years ago on AI.
00:01:13.215 --> 00:01:21.525
And that led me to deal more deeply with this, but also the classes that I teach where we talk about what it means to be human in different cultures.
00:01:22.387 --> 00:01:25.510
And so I started asking, AI is influencing our culture.
00:01:26.052 --> 00:01:26.292
How?
00:01:27.013 --> 00:01:27.933
Professor Hunt?
00:01:28.097 --> 00:01:37.230
I'd like you to summarize your greatest hopes for artificial intelligence as it is integrated into nearly every aspect of our world.
00:01:37.871 --> 00:01:45.063
What are the most positive outcomes for the AI tool and how will we and the planet benefit?
00:01:46.043 --> 00:02:05.956
I think the most positive tools are first those in science and medicine where AI tools and the basic technology behind AI is able to help us quickly advance in scientific experimentation, discovery of new drugs, potentially new energy sources, clean energy sources.
00:02:06.578 --> 00:02:07.759
There's a lot of potential there.
00:02:08.419 --> 00:02:19.657
There's also a lot of potential on the human side because AI does provide a way of having access to information and companionship for people who may not have it otherwise.
00:02:20.418 --> 00:02:20.938
You apply
00:02:20.979 --> 00:02:22.981
healthy skepticism to AI.
00:02:23.425 --> 00:02:27.030
which is evident in the title of your book and in other writings.
00:02:27.670 --> 00:02:42.128
We first met a few years ago when you and a student collaborated on an opinion piece about how chat GPT might impact higher education and eventually nearly all sectors of our world.
00:02:42.709 --> 00:02:43.330
Well, that's right.
00:02:43.349 --> 00:02:49.617
I'm not sure I'm an AI skeptic as much as I think we need to take seriously the dangers involved with AI.
00:02:50.145 --> 00:02:56.074
and particularly how we humans interact with AI and how it makes us think about ourselves as humans.
00:02:56.717 --> 00:03:05.510
That's the key part of my book, and the key thing I'm worried about is that if we treat a computer like a human, we may start treating ourselves like computers.
00:03:07.212 --> 00:03:19.991
Well, the Declaration of Independence states that we are endowed by our creator with certain unalienable rights, that among these are life, liberty, and the pursuit of happiness.
00:03:20.752 --> 00:03:24.777
Does AI enhance or unravel these lofty hopes?
00:03:25.437 --> 00:03:26.759
Well, it's a mixed bag.
00:03:27.439 --> 00:03:34.407
The possibility of AI giving us longer lives through medicine is very real and very hopeful.
00:03:35.348 --> 00:03:47.102
At the same time, it has a tremendous environmental cost, and there's a real threat to our human life on Earth and the other species if we cannot contain that environmental cost.
00:03:48.354 --> 00:03:52.920
When you talk about liberty, there I'm afraid it's more danger than anything else.
00:03:54.002 --> 00:04:24.281
Yes, it does liberate us by giving us new access to knowledge, new tools that we can use to do good things, but its ability to mimic humans and its ability to create deep fakes of videos and audio and writing really means that we are liable to be manipulated by people who have these tools at hand and be brought to places of not understanding what's really going on in our world because we're caught up in a world created by artificial intelligences.
00:04:25.122 --> 00:04:42.483
We've had some great AI conversations leading up to this podcast, but one of the most intriguing observations you've made is how we as humans tend to ascribe humanity on our AI tools and bots and how that kind of undermines our own existence.
00:04:43.064 --> 00:04:43.805
Please elaborate.
00:04:44.898 --> 00:04:47.240
It's long known that humans anthropomorphize.
00:04:47.420 --> 00:04:54.968
We see something that vaguely acts human or might be human, and we say, we treat it as if it's human.
00:04:55.028 --> 00:04:59.793
We ascribe to it certain characteristics that we would normally ascribe to ourselves.
00:05:00.495 --> 00:05:03.358
And that's a well-known tendency in us, and we all know it.
00:05:04.059 --> 00:05:13.490
The more realistic this form of false humanity becomes, the more like we are to anthropomorphize and treat it as a human.
00:05:14.017 --> 00:05:26.372
And I think the real danger here is not only that it reflects back to us a false view of our own humanity, but it begins to capture our attention and take us away from real human interactions with our fellow humans.
00:05:26.653 --> 00:05:29.697
It becomes a substitute for real human interactions.
00:05:30.338 --> 00:05:43.713
And you noted that some builders of robots like to make them look more like humans, and others want them to be totally mechanical creatures that may intimidate us.
00:05:44.065 --> 00:05:44.165
Well,
00:05:44.185 --> 00:05:44.567
that's right.
00:05:44.607 --> 00:05:48.009
I think a couple of different bets are being made by the makers of robots.
00:05:48.990 --> 00:05:54.317
One is that we'll be freaked out by them, as Isaac Asimov thought we might be in his book, I, Robot.
00:05:54.898 --> 00:06:00.863
And so they need to look like robots so that we don't ascribe too much humanity to them and we feel more comfortable with a robot.
00:06:01.584 --> 00:06:09.552
There's something called the uncanny valley, where you cross the uncanny valley and you get even more uncertain about what you're dealing with.
00:06:10.274 --> 00:06:20.925
But there's obviously other robot manufacturers, primarily in Asia, who really want to create a very, very life-human-like kind of android that we can relate to more humanly.
00:06:21.826 --> 00:06:31.336
I'm not sure which of these is going to freak us out more, but I have to say that I would prefer my robot to look like a robot, even if it's only got one eye and it's got mechanical fingers.
00:06:32.076 --> 00:06:32.557
Okay.
00:06:33.077 --> 00:06:44.802
Professor, you envision a kind of slippery slope for humankind as our contemporary social settings push us into greater interaction with human-like chatbots?
00:06:45.341 --> 00:06:45.942
Well, that's right.
00:06:46.023 --> 00:06:58.654
I think the danger is that, as I said earlier, the human-like interactions with these chatbots will lead us to exclude normal social interactions from our repertoire and will lead us away from each other.
00:06:58.694 --> 00:07:01.697
So I think that's a very key danger.
00:07:01.737 --> 00:07:07.584
And we may become engaged in communities that are almost entirely fake.
00:07:08.584 --> 00:07:09.505
So, for example...
00:07:09.954 --> 00:07:14.982
We already know that when we encounter someone on Facebook or Instagram, it may be a bot.
00:07:15.543 --> 00:07:19.048
It may be that we're having conversations with AIs and not our fellow humans.
00:07:19.928 --> 00:07:22.774
The same thing is true in some of the games that are being played now.
00:07:23.355 --> 00:07:27.440
AI robots are game players pretending to be human.
00:07:28.423 --> 00:07:31.286
So I think that's the big social danger there.
00:07:31.307 --> 00:07:40.153
And it's the reason that And I would say it's enhanced, too, by the fact that we're used to dealing with screens, two-dimensional representations of reality.
00:07:41.113 --> 00:07:53.848
And in a screen world, in a screen age, it's all the easier to be fooled by that bot, which would be hard-pressed to fool us in the real 3D physical world.
00:07:54.569 --> 00:07:56.430
I think this is the key danger to us.
00:07:56.911 --> 00:07:59.454
I do want to stress, there are possibilities here.
00:07:59.795 --> 00:08:00.995
There are people who are lonely.
00:08:01.016 --> 00:08:14.110
And A short cure for loneliness may be a chatbot, but the long cure for loneliness is that we as human beings treat each other as human beings and come into each other's presence and care for each other.
00:08:14.790 --> 00:08:15.190
Yikes.
00:08:15.891 --> 00:08:25.661
So what are some pushback mechanisms we can use to counteract this potentially slow descent into AI subjugation?
00:08:26.201 --> 00:08:29.505
Well, I think the main one is that we be humans with humans.
00:08:30.081 --> 00:08:31.944
There's talk now about third spaces.
00:08:32.024 --> 00:08:53.394
Third spaces are places away from our work, away from our family, away from our school, where we just meet human beings as human beings and have a chance to interact with each other without all the mediation of technology, or at least without the mediation of chatbots in the room.
00:08:54.195 --> 00:08:56.138
And third spaces like that are going to be important.
00:08:56.638 --> 00:09:01.923
Schools, clubs, churches, sailboat clubs, running clubs.
00:09:02.725 --> 00:09:08.070
All these places are places where we can be with humans and defend our own humanity.
00:09:08.610 --> 00:09:14.235
But there's another important part of that too, which is we need to demystify the AI.
00:09:14.995 --> 00:09:34.778
When we hear a chatbot that sounds incredibly realistic, and there's a new one that is most realistic I've ever heard, we need to remind ourselves that it's a bunch of microprocessors running a bunch of algorithmic procedures that generate language on a basis of probabilities.
00:09:35.600 --> 00:09:36.461
It isn't conscious.
00:09:37.442 --> 00:09:38.503
It doesn't care about us.
00:09:39.724 --> 00:09:41.706
It does a great mimic of caring about us.
00:09:42.047 --> 00:09:45.210
It's not conscious, doesn't care about us, doesn't know we exist.
00:09:46.250 --> 00:09:49.414
And by the way, we don't have any influence on it.
00:09:49.434 --> 00:09:52.457
And I think this is really important.
00:09:54.240 --> 00:09:59.426
When we talk to a chat bot, it appears to listen to us and respond to our words.
00:10:00.802 --> 00:10:07.251
But by the very way it is made and trained, it cannot internalize anything about us.
00:10:08.494 --> 00:10:09.434
It can change us.
00:10:10.375 --> 00:10:12.759
We cannot change it at all.
00:10:13.660 --> 00:10:14.722
It can change our minds.
00:10:15.865 --> 00:10:17.547
We cannot change its mind.
00:10:18.668 --> 00:10:24.817
And real human interactions depend on us forming and shaping each other, not just being formed and shaped.
00:10:25.418 --> 00:10:42.596
When we meet someone who manipulates us but doesn't listen to us, who wants to change us but will not be changed by us and our feelings, then we rightly regard that person as a predator or a narcissist.
00:10:43.216 --> 00:10:44.357
We wouldn't interact with them.
00:10:45.841 --> 00:10:52.230
Well, we need to remind ourselves how very useful chatbots are, and they're extremely useful, and I use them all the time.
00:10:52.269 --> 00:10:54.974
I'm not doing anything for them.
00:10:55.895 --> 00:10:57.097
They don't care that I exist.
00:10:57.601 --> 00:11:03.129
One second after I quit talking, they forget everything I've said or they store it in some kind of long-term memory to dredge it up later.
00:11:03.149 --> 00:11:14.644
And I'm missing, therefore, something critical about this part of what it means for me to be human, which is to help form another person's humanity, even as they form mine.
00:11:15.525 --> 00:11:25.219
So you've mentioned a lot of the myths about AI tools and our interactions with them, but this isn't a myth.
00:11:25.779 --> 00:11:34.811
I think many of us Users are led to believe AI is looking out for us, that it has ethics and a moral compass.
00:11:36.113 --> 00:11:41.080
Something tells me, like the title of your book, that you might question that.
00:11:42.000 --> 00:11:42.662
Indeed I do.
00:11:43.602 --> 00:11:46.086
It has no ethic because it has no consciousness.
00:11:47.027 --> 00:11:50.532
It can't see a bigger picture, which is kind of the essence of ethics.
00:11:51.673 --> 00:11:55.177
It, if given a question or a statement...
00:11:55.778 --> 00:11:58.902
will generate a set of processes that create a response.
00:11:59.942 --> 00:12:00.323
That's all.
00:12:01.445 --> 00:12:02.105
That's all it does.
00:12:03.267 --> 00:12:08.413
If that response happens to be hurtful or negative, it won't even realize it.
00:12:09.634 --> 00:12:21.590
Now, AI can be trained in such a way that it limits the negativity of its responses, although many AI vendors won't do that.
00:12:22.529 --> 00:12:24.652
But that's not the same as having an ethic.
00:12:25.253 --> 00:12:29.317
The classic case of this is an old idea from 20 years ago or so.
00:12:30.638 --> 00:12:45.833
It's the problem of what happens if you ask an AI to do something and it goes about doing it, but in order to do what you ask it to do, it begins to kill things or destroy things because it has no big picture of the consequences of its actions.
00:12:46.234 --> 00:12:47.754
You tell it to make paper clips.
00:12:47.975 --> 00:12:49.937
This is Nate Boston's classic thing.
00:12:50.529 --> 00:12:55.535
And it's going to make paperclips as efficiently as possible, even if it means to turn humans into paperclips.
00:12:56.576 --> 00:12:59.681
AI just doesn't have that big picture.
00:13:00.162 --> 00:13:07.490
It doesn't step outside itself and ask about the consequences of its actions, or if it's been trained to do so.
00:13:07.530 --> 00:13:11.895
It's extraordinarily limited compared to human beings.
00:13:12.636 --> 00:13:15.961
Professor Hunt, we've discussed many AI issues.
00:13:16.802 --> 00:13:24.714
But I'd like to take this opportunity to invite you to bring up other AI matters you'd like our listeners to be aware of.
00:13:25.274 --> 00:13:29.640
These can be positives or problematic AI insights.
00:13:30.381 --> 00:13:38.254
What is top of my mind right now is the way in which AI is going to disrupt and reshape our economy.
00:13:39.475 --> 00:13:46.145
And of course, in doing that, it disrupts and reshapes everything related to the economy, including higher education.
00:13:47.394 --> 00:13:50.356
Now, I'll give two quick examples of how this is going to happen.
00:13:51.278 --> 00:14:03.090
One is that although currently most AI agents or agentic AIs, these are AIs that can operate somewhat autonomously, most of them now are being used to help people do their work better.
00:14:04.711 --> 00:14:07.114
But they are getting good enough to replace people.
00:14:08.816 --> 00:14:14.701
And I don't think there's any question that for people who run a business, they will replace people with AI.
00:14:16.129 --> 00:14:18.011
That, in fact, they've already said they will.
00:14:19.053 --> 00:14:33.211
Even someone like Dario Amodi of Anthropic, who is one of the most ethical developers of AI, has said they're going to be hiring fewer software engineers and letting old ones move and retire.
00:14:35.072 --> 00:14:36.434
The same thing is happening in law.
00:14:37.655 --> 00:14:45.426
One of the largest law firms in the world has said, we're going to hire many fewer junior lawyers because we have agent AIs that can take that place.
00:14:46.210 --> 00:14:49.835
We're going to see that happen across the economy at an accelerating pace.
00:14:50.576 --> 00:14:52.097
And so that's going to cause disruption.
00:14:52.879 --> 00:14:55.763
Typically, new technologies have caused this disruption.
00:14:56.725 --> 00:14:58.746
Frequently, new jobs have occurred.
00:15:00.168 --> 00:15:02.673
But we've got to be aware that the disruption is going to happen.
00:15:03.693 --> 00:15:07.940
This is top of mind now because we're entering into troubled economic waters already.
00:15:08.821 --> 00:15:12.025
Increased unemployment is going to be very problematic for us.
00:15:12.405 --> 00:15:14.990
And increasing cases of people who cannot be employed.
00:15:16.706 --> 00:15:20.370
This is something big for us in higher education that we have to take seriously.
00:15:21.892 --> 00:15:26.557
Are we preparing people for jobs that will be replaced by AI?
00:15:28.679 --> 00:15:30.783
And I think the answer in some cases is yes.
00:15:32.884 --> 00:15:40.494
That doesn't mean that our students cannot find other ways to use their brilliance and creativity, and we've got a brilliant and creative batch here at SMU.
00:15:41.730 --> 00:15:48.860
But it may well be that the jobs that they thought they were going to have when they came in as freshmen don't exist when they graduate as seniors.
00:15:50.283 --> 00:15:52.066
And that's something we have to take seriously.
00:15:52.105 --> 00:15:54.048
That's an ethical problem for us.
00:15:55.129 --> 00:16:02.240
Professor Hunt, a little AI bird has informed me that you're gearing up to do a podcast series of your own.
00:16:03.162 --> 00:16:06.407
How many episodes and what kind of topics do you wish to cover?
00:16:07.808 --> 00:16:15.033
Well, yes, I'm going to do a podcast, and it's going to be on– robertihunt.com, probably starting in a couple of weeks.
00:16:15.673 --> 00:16:17.535
I have not said a number of episodes.
00:16:18.116 --> 00:16:36.595
What I'm going to try to do is every couple of weeks take up the latest news in AI and ask how that specifically addresses or challenges what's happening with us as humans who are defending our own humanity, and therefore try to keep the focus on here's something that's in our world.
00:16:37.134 --> 00:16:39.738
We need to know it, understand it, and what it's doing.
00:16:40.322 --> 00:16:41.803
But we also need to know ourselves.
00:16:42.644 --> 00:16:47.412
And as we know ourselves, continue our own quest to be the kind of humans we ought to be.
00:16:47.432 --> 00:16:47.451
A
00:16:48.313 --> 00:16:49.174
final question.
00:16:49.475 --> 00:16:54.302
Are enough of us paying attention to this, or are we oblivious?
00:16:55.062 --> 00:16:59.929
I'm not sure that people are oblivious, but not nearly enough are paying attention across the board.
00:16:59.970 --> 00:17:05.919
I think in business and industry where AI has a direct impact, there's a lot of attention being paid.
00:17:06.740 --> 00:17:07.842
But I think a lot of people...
00:17:08.642 --> 00:17:13.169
are in areas where AI is influencing their world and they don't know it.
00:17:13.769 --> 00:17:15.152
And so they're not paying attention.
00:17:15.752 --> 00:17:17.634
And of course, we have a lot of other distractions.
00:17:18.155 --> 00:17:19.778
But I think we should be paying attention.
00:17:20.619 --> 00:17:29.153
I'm one of those who believes that we're going to see in the next 12 to 24 months pretty serious disruption of some of our social systems by AI.
00:17:30.114 --> 00:17:30.954
If I'm wrong, fine.
00:17:31.777 --> 00:17:34.160
But if I'm not wrong, then we need to start paying attention.
00:17:34.921 --> 00:17:37.885
Professor Hunt, thank you so much for being with us today.
00:17:38.402 --> 00:17:42.588
and sharing your hopes and concerns regarding artificial intelligence.
00:17:43.691 --> 00:17:58.996
I'd also like to thank my colleague Stephen Fasaro for his technical support and the SMU Office of Engaged Learning, which makes the SMU Fondren Library Podcasting Studio available to us for our recordings.
00:17:59.758 --> 00:18:00.558
Until next time.
00:18:01.220 --> 00:18:02.461
Thank you very much, Bob.
00:00:05.921 --> 00:00:08.826
Welcome to the SMU Perspectives Podcast.
00:00:09.307 --> 00:00:12.089
I am Robert Ehlert, the SMU Commentary Editor.
00:00:12.109 --> 00:00:21.603
Our guest today is Robert Hunt, Director of Global Theological Education at the SMU Perkins School of Theology.
00:00:21.663 --> 00:00:32.398
Professor Hunt is the author of several books, but his latest delves into the intersection of artificial intelligence and the human species.
00:00:32.417 --> 00:00:35.502
It's called All Brain and No Soul.
00:00:36.130 --> 00:00:38.954
Real Humanity in an AI Age.
00:00:39.935 --> 00:00:50.572
Before we dive into that subject, Professor Hunt, elaborate how someone with a theology background began probing AI and its impact on our society.
00:00:51.292 --> 00:00:52.213
Thank you very much, Bob.
00:00:52.575 --> 00:00:53.155
A couple of things.
00:00:53.256 --> 00:00:57.802
First, I did actually have a long background in computers and computer science.
00:00:57.923 --> 00:01:00.225
I started as a computer programmer in university.
00:01:00.646 --> 00:01:02.289
Then I switched to history and theology.
00:01:02.722 --> 00:01:06.126
And so I followed this very closely over the course of my career.
00:01:06.947 --> 00:01:13.174
What brought this more immediately to mind was a conference we had at SMU about six years ago on AI.
00:01:13.215 --> 00:01:21.525
And that led me to deal more deeply with this, but also the classes that I teach where we talk about what it means to be human in different cultures.
00:01:22.387 --> 00:01:25.510
And so I started asking, AI is influencing our culture.
00:01:26.052 --> 00:01:26.292
How?
00:01:27.013 --> 00:01:27.933
Professor Hunt?
00:01:28.097 --> 00:01:37.230
I'd like you to summarize your greatest hopes for artificial intelligence as it is integrated into nearly every aspect of our world.
00:01:37.871 --> 00:01:45.063
What are the most positive outcomes for the AI tool and how will we and the planet benefit?
00:01:46.043 --> 00:02:05.956
I think the most positive tools are first those in science and medicine where AI tools and the basic technology behind AI is able to help us quickly advance in scientific experimentation, discovery of new drugs, potentially new energy sources, clean energy sources.
00:02:06.578 --> 00:02:07.759
There's a lot of potential there.
00:02:08.419 --> 00:02:19.657
There's also a lot of potential on the human side because AI does provide a way of having access to information and companionship for people who may not have it otherwise.
00:02:20.418 --> 00:02:20.938
You apply
00:02:20.979 --> 00:02:22.981
healthy skepticism to AI.
00:02:23.425 --> 00:02:27.030
which is evident in the title of your book and in other writings.
00:02:27.670 --> 00:02:42.128
We first met a few years ago when you and a student collaborated on an opinion piece about how chat GPT might impact higher education and eventually nearly all sectors of our world.
00:02:42.709 --> 00:02:43.330
Well, that's right.
00:02:43.349 --> 00:02:49.617
I'm not sure I'm an AI skeptic as much as I think we need to take seriously the dangers involved with AI.
00:02:50.145 --> 00:02:56.074
and particularly how we humans interact with AI and how it makes us think about ourselves as humans.
00:02:56.717 --> 00:03:05.510
That's the key part of my book, and the key thing I'm worried about is that if we treat a computer like a human, we may start treating ourselves like computers.
00:03:07.212 --> 00:03:19.991
Well, the Declaration of Independence states that we are endowed by our creator with certain unalienable rights, that among these are life, liberty, and the pursuit of happiness.
00:03:20.752 --> 00:03:24.777
Does AI enhance or unravel these lofty hopes?
00:03:25.437 --> 00:03:26.759
Well, it's a mixed bag.
00:03:27.439 --> 00:03:34.407
The possibility of AI giving us longer lives through medicine is very real and very hopeful.
00:03:35.348 --> 00:03:47.102
At the same time, it has a tremendous environmental cost, and there's a real threat to our human life on Earth and the other species if we cannot contain that environmental cost.
00:03:48.354 --> 00:03:52.920
When you talk about liberty, there I'm afraid it's more danger than anything else.
00:03:54.002 --> 00:04:24.281
Yes, it does liberate us by giving us new access to knowledge, new tools that we can use to do good things, but its ability to mimic humans and its ability to create deep fakes of videos and audio and writing really means that we are liable to be manipulated by people who have these tools at hand and be brought to places of not understanding what's really going on in our world because we're caught up in a world created by artificial intelligences.
00:04:25.122 --> 00:04:42.483
We've had some great AI conversations leading up to this podcast, but one of the most intriguing observations you've made is how we as humans tend to ascribe humanity on our AI tools and bots and how that kind of undermines our own existence.
00:04:43.064 --> 00:04:43.805
Please elaborate.
00:04:44.898 --> 00:04:47.240
It's long known that humans anthropomorphize.
00:04:47.420 --> 00:04:54.968
We see something that vaguely acts human or might be human, and we say, we treat it as if it's human.
00:04:55.028 --> 00:04:59.793
We ascribe to it certain characteristics that we would normally ascribe to ourselves.
00:05:00.495 --> 00:05:03.358
And that's a well-known tendency in us, and we all know it.
00:05:04.059 --> 00:05:13.490
The more realistic this form of false humanity becomes, the more like we are to anthropomorphize and treat it as a human.
00:05:14.017 --> 00:05:26.372
And I think the real danger here is not only that it reflects back to us a false view of our own humanity, but it begins to capture our attention and take us away from real human interactions with our fellow humans.
00:05:26.653 --> 00:05:29.697
It becomes a substitute for real human interactions.
00:05:30.338 --> 00:05:43.713
And you noted that some builders of robots like to make them look more like humans, and others want them to be totally mechanical creatures that may intimidate us.
00:05:44.065 --> 00:05:44.165
Well,
00:05:44.185 --> 00:05:44.567
that's right.
00:05:44.607 --> 00:05:48.009
I think a couple of different bets are being made by the makers of robots.
00:05:48.990 --> 00:05:54.317
One is that we'll be freaked out by them, as Isaac Asimov thought we might be in his book, I, Robot.
00:05:54.898 --> 00:06:00.863
And so they need to look like robots so that we don't ascribe too much humanity to them and we feel more comfortable with a robot.
00:06:01.584 --> 00:06:09.552
There's something called the uncanny valley, where you cross the uncanny valley and you get even more uncertain about what you're dealing with.
00:06:10.274 --> 00:06:20.925
But there's obviously other robot manufacturers, primarily in Asia, who really want to create a very, very life-human-like kind of android that we can relate to more humanly.
00:06:21.826 --> 00:06:31.336
I'm not sure which of these is going to freak us out more, but I have to say that I would prefer my robot to look like a robot, even if it's only got one eye and it's got mechanical fingers.
00:06:32.076 --> 00:06:32.557
Okay.
00:06:33.077 --> 00:06:44.802
Professor, you envision a kind of slippery slope for humankind as our contemporary social settings push us into greater interaction with human-like chatbots?
00:06:45.341 --> 00:06:45.942
Well, that's right.
00:06:46.023 --> 00:06:58.654
I think the danger is that, as I said earlier, the human-like interactions with these chatbots will lead us to exclude normal social interactions from our repertoire and will lead us away from each other.
00:06:58.694 --> 00:07:01.697
So I think that's a very key danger.
00:07:01.737 --> 00:07:07.584
And we may become engaged in communities that are almost entirely fake.
00:07:08.584 --> 00:07:09.505
So, for example...
00:07:09.954 --> 00:07:14.982
We already know that when we encounter someone on Facebook or Instagram, it may be a bot.
00:07:15.543 --> 00:07:19.048
It may be that we're having conversations with AIs and not our fellow humans.
00:07:19.928 --> 00:07:22.774
The same thing is true in some of the games that are being played now.
00:07:23.355 --> 00:07:27.440
AI robots are game players pretending to be human.
00:07:28.423 --> 00:07:31.286
So I think that's the big social danger there.
00:07:31.307 --> 00:07:40.153
And it's the reason that And I would say it's enhanced, too, by the fact that we're used to dealing with screens, two-dimensional representations of reality.
00:07:41.113 --> 00:07:53.848
And in a screen world, in a screen age, it's all the easier to be fooled by that bot, which would be hard-pressed to fool us in the real 3D physical world.
00:07:54.569 --> 00:07:56.430
I think this is the key danger to us.
00:07:56.911 --> 00:07:59.454
I do want to stress, there are possibilities here.
00:07:59.795 --> 00:08:00.995
There are people who are lonely.
00:08:01.016 --> 00:08:14.110
And A short cure for loneliness may be a chatbot, but the long cure for loneliness is that we as human beings treat each other as human beings and come into each other's presence and care for each other.
00:08:14.790 --> 00:08:15.190
Yikes.
00:08:15.891 --> 00:08:25.661
So what are some pushback mechanisms we can use to counteract this potentially slow descent into AI subjugation?
00:08:26.201 --> 00:08:29.505
Well, I think the main one is that we be humans with humans.
00:08:30.081 --> 00:08:31.944
There's talk now about third spaces.
00:08:32.024 --> 00:08:53.394
Third spaces are places away from our work, away from our family, away from our school, where we just meet human beings as human beings and have a chance to interact with each other without all the mediation of technology, or at least without the mediation of chatbots in the room.
00:08:54.195 --> 00:08:56.138
And third spaces like that are going to be important.
00:08:56.638 --> 00:09:01.923
Schools, clubs, churches, sailboat clubs, running clubs.
00:09:02.725 --> 00:09:08.070
All these places are places where we can be with humans and defend our own humanity.
00:09:08.610 --> 00:09:14.235
But there's another important part of that too, which is we need to demystify the AI.
00:09:14.995 --> 00:09:34.778
When we hear a chatbot that sounds incredibly realistic, and there's a new one that is most realistic I've ever heard, we need to remind ourselves that it's a bunch of microprocessors running a bunch of algorithmic procedures that generate language on a basis of probabilities.
00:09:35.600 --> 00:09:36.461
It isn't conscious.
00:09:37.442 --> 00:09:38.503
It doesn't care about us.
00:09:39.724 --> 00:09:41.706
It does a great mimic of caring about us.
00:09:42.047 --> 00:09:45.210
It's not conscious, doesn't care about us, doesn't know we exist.
00:09:46.250 --> 00:09:49.414
And by the way, we don't have any influence on it.
00:09:49.434 --> 00:09:52.457
And I think this is really important.
00:09:54.240 --> 00:09:59.426
When we talk to a chat bot, it appears to listen to us and respond to our words.
00:10:00.802 --> 00:10:07.251
But by the very way it is made and trained, it cannot internalize anything about us.
00:10:08.494 --> 00:10:09.434
It can change us.
00:10:10.375 --> 00:10:12.759
We cannot change it at all.
00:10:13.660 --> 00:10:14.722
It can change our minds.
00:10:15.865 --> 00:10:17.547
We cannot change its mind.
00:10:18.668 --> 00:10:24.817
And real human interactions depend on us forming and shaping each other, not just being formed and shaped.
00:10:25.418 --> 00:10:42.596
When we meet someone who manipulates us but doesn't listen to us, who wants to change us but will not be changed by us and our feelings, then we rightly regard that person as a predator or a narcissist.
00:10:43.216 --> 00:10:44.357
We wouldn't interact with them.
00:10:45.841 --> 00:10:52.230
Well, we need to remind ourselves how very useful chatbots are, and they're extremely useful, and I use them all the time.
00:10:52.269 --> 00:10:54.974
I'm not doing anything for them.
00:10:55.895 --> 00:10:57.097
They don't care that I exist.
00:10:57.601 --> 00:11:03.129
One second after I quit talking, they forget everything I've said or they store it in some kind of long-term memory to dredge it up later.
00:11:03.149 --> 00:11:14.644
And I'm missing, therefore, something critical about this part of what it means for me to be human, which is to help form another person's humanity, even as they form mine.
00:11:15.525 --> 00:11:25.219
So you've mentioned a lot of the myths about AI tools and our interactions with them, but this isn't a myth.
00:11:25.779 --> 00:11:34.811
I think many of us Users are led to believe AI is looking out for us, that it has ethics and a moral compass.
00:11:36.113 --> 00:11:41.080
Something tells me, like the title of your book, that you might question that.
00:11:42.000 --> 00:11:42.662
Indeed I do.
00:11:43.602 --> 00:11:46.086
It has no ethic because it has no consciousness.
00:11:47.027 --> 00:11:50.532
It can't see a bigger picture, which is kind of the essence of ethics.
00:11:51.673 --> 00:11:55.177
It, if given a question or a statement...
00:11:55.778 --> 00:11:58.902
will generate a set of processes that create a response.
00:11:59.942 --> 00:12:00.323
That's all.
00:12:01.445 --> 00:12:02.105
That's all it does.
00:12:03.267 --> 00:12:08.413
If that response happens to be hurtful or negative, it won't even realize it.
00:12:09.634 --> 00:12:21.590
Now, AI can be trained in such a way that it limits the negativity of its responses, although many AI vendors won't do that.
00:12:22.529 --> 00:12:24.652
But that's not the same as having an ethic.
00:12:25.253 --> 00:12:29.317
The classic case of this is an old idea from 20 years ago or so.
00:12:30.638 --> 00:12:45.833
It's the problem of what happens if you ask an AI to do something and it goes about doing it, but in order to do what you ask it to do, it begins to kill things or destroy things because it has no big picture of the consequences of its actions.
00:12:46.234 --> 00:12:47.754
You tell it to make paper clips.
00:12:47.975 --> 00:12:49.937
This is Nate Boston's classic thing.
00:12:50.529 --> 00:12:55.535
And it's going to make paperclips as efficiently as possible, even if it means to turn humans into paperclips.
00:12:56.576 --> 00:12:59.681
AI just doesn't have that big picture.
00:13:00.162 --> 00:13:07.490
It doesn't step outside itself and ask about the consequences of its actions, or if it's been trained to do so.
00:13:07.530 --> 00:13:11.895
It's extraordinarily limited compared to human beings.
00:13:12.636 --> 00:13:15.961
Professor Hunt, we've discussed many AI issues.
00:13:16.802 --> 00:13:24.714
But I'd like to take this opportunity to invite you to bring up other AI matters you'd like our listeners to be aware of.
00:13:25.274 --> 00:13:29.640
These can be positives or problematic AI insights.
00:13:30.381 --> 00:13:38.254
What is top of my mind right now is the way in which AI is going to disrupt and reshape our economy.
00:13:39.475 --> 00:13:46.145
And of course, in doing that, it disrupts and reshapes everything related to the economy, including higher education.
00:13:47.394 --> 00:13:50.356
Now, I'll give two quick examples of how this is going to happen.
00:13:51.278 --> 00:14:03.090
One is that although currently most AI agents or agentic AIs, these are AIs that can operate somewhat autonomously, most of them now are being used to help people do their work better.
00:14:04.711 --> 00:14:07.114
But they are getting good enough to replace people.
00:14:08.816 --> 00:14:14.701
And I don't think there's any question that for people who run a business, they will replace people with AI.
00:14:16.129 --> 00:14:18.011
That, in fact, they've already said they will.
00:14:19.053 --> 00:14:33.211
Even someone like Dario Amodi of Anthropic, who is one of the most ethical developers of AI, has said they're going to be hiring fewer software engineers and letting old ones move and retire.
00:14:35.072 --> 00:14:36.434
The same thing is happening in law.
00:14:37.655 --> 00:14:45.426
One of the largest law firms in the world has said, we're going to hire many fewer junior lawyers because we have agent AIs that can take that place.
00:14:46.210 --> 00:14:49.835
We're going to see that happen across the economy at an accelerating pace.
00:14:50.576 --> 00:14:52.097
And so that's going to cause disruption.
00:14:52.879 --> 00:14:55.763
Typically, new technologies have caused this disruption.
00:14:56.725 --> 00:14:58.746
Frequently, new jobs have occurred.
00:15:00.168 --> 00:15:02.673
But we've got to be aware that the disruption is going to happen.
00:15:03.693 --> 00:15:07.940
This is top of mind now because we're entering into troubled economic waters already.
00:15:08.821 --> 00:15:12.025
Increased unemployment is going to be very problematic for us.
00:15:12.405 --> 00:15:14.990
And increasing cases of people who cannot be employed.
00:15:16.706 --> 00:15:20.370
This is something big for us in higher education that we have to take seriously.
00:15:21.892 --> 00:15:26.557
Are we preparing people for jobs that will be replaced by AI?
00:15:28.679 --> 00:15:30.783
And I think the answer in some cases is yes.
00:15:32.884 --> 00:15:40.494
That doesn't mean that our students cannot find other ways to use their brilliance and creativity, and we've got a brilliant and creative batch here at SMU.
00:15:41.730 --> 00:15:48.860
But it may well be that the jobs that they thought they were going to have when they came in as freshmen don't exist when they graduate as seniors.
00:15:50.283 --> 00:15:52.066
And that's something we have to take seriously.
00:15:52.105 --> 00:15:54.048
That's an ethical problem for us.
00:15:55.129 --> 00:16:02.240
Professor Hunt, a little AI bird has informed me that you're gearing up to do a podcast series of your own.
00:16:03.162 --> 00:16:06.407
How many episodes and what kind of topics do you wish to cover?
00:16:07.808 --> 00:16:15.033
Well, yes, I'm going to do a podcast, and it's going to be on– robertihunt.com, probably starting in a couple of weeks.
00:16:15.673 --> 00:16:17.535
I have not said a number of episodes.
00:16:18.116 --> 00:16:36.595
What I'm going to try to do is every couple of weeks take up the latest news in AI and ask how that specifically addresses or challenges what's happening with us as humans who are defending our own humanity, and therefore try to keep the focus on here's something that's in our world.
00:16:37.134 --> 00:16:39.738
We need to know it, understand it, and what it's doing.
00:16:40.322 --> 00:16:41.803
But we also need to know ourselves.
00:16:42.644 --> 00:16:47.412
And as we know ourselves, continue our own quest to be the kind of humans we ought to be.
00:16:47.432 --> 00:16:47.451
A
00:16:48.313 --> 00:16:49.174
final question.
00:16:49.475 --> 00:16:54.302
Are enough of us paying attention to this, or are we oblivious?
00:16:55.062 --> 00:16:59.929
I'm not sure that people are oblivious, but not nearly enough are paying attention across the board.
00:16:59.970 --> 00:17:05.919
I think in business and industry where AI has a direct impact, there's a lot of attention being paid.
00:17:06.740 --> 00:17:07.842
But I think a lot of people...
00:17:08.642 --> 00:17:13.169
are in areas where AI is influencing their world and they don't know it.
00:17:13.769 --> 00:17:15.152
And so they're not paying attention.
00:17:15.752 --> 00:17:17.634
And of course, we have a lot of other distractions.
00:17:18.155 --> 00:17:19.778
But I think we should be paying attention.
00:17:20.619 --> 00:17:29.153
I'm one of those who believes that we're going to see in the next 12 to 24 months pretty serious disruption of some of our social systems by AI.
00:17:30.114 --> 00:17:30.954
If I'm wrong, fine.
00:17:31.777 --> 00:17:34.160
But if I'm not wrong, then we need to start paying attention.
00:17:34.921 --> 00:17:37.885
Professor Hunt, thank you so much for being with us today.
00:17:38.402 --> 00:17:42.588
and sharing your hopes and concerns regarding artificial intelligence.
00:17:43.691 --> 00:17:58.996
I'd also like to thank my colleague Stephen Fasaro for his technical support and the SMU Office of Engaged Learning, which makes the SMU Fondren Library Podcasting Studio available to us for our recordings.
00:17:59.758 --> 00:18:00.558
Until next time.
00:18:01.220 --> 00:18:02.461
Thank you very much, Bob.
