The Philosophy of AI: Examining the Ethical Implications with David Danks (Part One)

Forbes Books presents the Startup Science Podcast with Gregory Shepard, brought to you by startupscience.io. Greg is an entrepreneur who has built and sold 12 businesses. He's a recipient of 4 private equity awards and is a featured TEDx speaker and Forbes author. Here's Gregory Shepard.

My guest today is David Dankes, a professor of data science and philosophy at the University of California, San Diego. His research explores the intersection of philosophy, cognitive science, and machine learning, tackling complex interdisciplinary problems. He's examined the ethical, psychological, and policy implications of AI robotics in various domains like transportation and healthcare. David, I'm excited to have you on the podcast, and I'd love it if you could start by talking a little bit about all of the incredible work you've done Bridging these different fields with your brilliant mind.

Well, thank you for that very kind introduction. So I'm here at UC San Diego, and I do everything from actually build AI systems, primarily with biomedical data, through to thinking about the ethics and policy of AI. On top of that, a lot of issues around philosophy more generally. So what can we possibly know about the world? How do we come to know about the world? What's the nature of science? I'm one of these people that I've discovered I don't actually have any one big problem that I'm trying to solve. Many academics do. Instead, I'm just kind of interested in everything and lucky enough to be in a position that I get to pursue all of these different ideas that might come to my mind.

So when it comes to AI and ethics and philosophy, it seems like philosophy and this new realm we're entering with AI, would— there would be a lot of benefits to understanding philosophy, right? Like, how do you look at the way we're using AI or what the future holds for AI, that sort of stuff? Do you have any perspectives on that?

Completely agree with you that philosophy is invaluable in our current time, you know, milieu. Um, I think that's not always obvious to people, in part because I think many people misunderstand the nature of philosophy. So I think philosophy is sometimes thought of as, you know, it's either the navel-gazing that occurs in a dorm room at midnight, or it's the province of a bunch of dead white guys from Greece onward. But what philosophy is really about is it's about asking the foundational questions. It's about asking those questions that in most of the parts of our lives, we don't have the time or even the opportunity and space to stop and ask them, you know, what is the nature of knowledge? What is the good life? And so when we find ourselves in a time of significant disruption, I think it's all the more important that we're able to go back to those foundations, to have a firm ground, at least about our own values, what we think is important, what we're trying to achieve in our lives. And philosophy is ultimately about giving people the skills and experience To be able to ask those questions about their own foundations.

Do you have, or are you still working on, a philosophy of AI? Right? Like, I think about it like this. It's moving so fast. It's been around for a long time, but it just hit this point where it's like really, really moving fast, and people are starting to apply it in all these different areas. And when you think about the philosophy of AI, are you thinking about it in terms of You know, it should be used for this and not used for that, or we can expect this to happen in the future, you know, and that sort of stuff. I'm just coming off of reading Sapiens, so I'd like to hear about that.

So a lot of my own work in philosophy and AI has focused on understanding the ways in which our values get implemented, advanced, almost kind of expressed through the AI systems that we build. And I think a lot of people think of AI as just code. It's just the ones and zeros or it's math, but I'm much more interested in the way that all of the different design choices, the choices made by the developers and creators of AI systems, all the ways that those choices then lead to the systems being designed for particular purposes, that they benefit some people and not others, that they're good in some contexts and not others. And so really trying to connect the values of the people creating the AI, the people using the AI, connecting those values with the values that matter to us in our lives. So I worry that a lot of the AI systems that are being built these days, they are benefiting just a few people and not others. So this means that I end up not thinking all that much about, you know, the definitional question, what counts as AI, what doesn't count as AI. Or thinking about very abstract questions about what kinds of AIs might exist in 30 years. And I'm much more trying to think about, you know, the, the, my neighbor who's a retired lawyer and is trying to figure out, should she be freaked out about AI? Or the person who's worried that AI is going to take their job. How do we think about how to create AI that empowers people rather than just building things that are Um, seem cool to me but might end up having a lot of harmful side effects.

Yeah, so let me ask you about this. I was talking to a pretty smart fellow in London a few weeks ago. What he said is, he said, listen, everybody was afraid of losing their jobs because of robotics, and they did, but then they got re-employed. And AI, on the other hand, is the other end of the spectrum, right? So you have the robotics, which is affected the blue-collar workers. And then you have AI, which is going to affect the white-collar workers, right? Like doctors, lawyers, things like AI is just going to do a better job than they could do in diagnosing or drafting or whatever it is. When you think about that, do you feel like there is going to be a point at which people aren't working as much? Right? So this guy said, he said, listen, you know, it's not like people aren't going to work, they're just going to work 4 hours instead of 12 hours or whatever. to do the same thing and get paid the same money. Do you feel like that's a future that, in terms of an optimistic outlook, do you feel like that's a future that is possible, or what are your thoughts on that?

So I think that's a future that is possible, but it's not going to happen without really concerted effort. Somewhat famously, John Maynard Keynes, the economist, thought that the steam engine and industrialization was going to mean that by the early 20th century, everybody would be working 15 to 20 hours and have lives of leisure. Obviously, that's not how it turned out. And I think we're facing a similar sort of choice here with AI. Do we take advantage of the massive productivity gains that we can achieve through AI to reshape in certain ways, quite frankly, our fundamental society? Do we have things like universal basic income that enable people to step away from work and pursue their passions, which for many people, passion is not the same as work. That's the part at which I start to be a little bit less optimistic because I think right now it's not clear to me that there is the political will, the political direction, the political coherence to be able to get us towards that better future. I think it's possible, but unfortunately not probable. In fact, I worry in many ways that the risks are much more on the downside than on the upside. In the sense that right now, you know, these past technological changes, even robotics, but if we go back earlier to, you know, cars replacing horses and these kinds of things, as you said, people have lost jobs, but they've always been able to get new jobs afterwards. And so the challenge though is that there's a period of disruption in the middle. And historically that's been 10+ years. It takes a long time for a new technology to come into widespread use in large part because technologies usually are physical things. So it just, you have to build more of them, right? So it takes time to convert all factories into being primarily run by robots because we have to build a lot of robots to do that. AI costs almost nothing to duplicate it. And so I think that the speed of adoption can be much faster with AI, which also means the speed of disruption and dislocation of workers is going to be much faster with AI. And I don't know if our society can handle suddenly having 20% unemployment. If it's not spread out over 10 years, I worry that there's actually a much bigger risk of significant disruption socially rather than the positive upside possibility.

So when you think about this, so I'm going to sort of move my way into entrepreneurship and startups and stuff like that. When you think about this, you think about industrial transfer right now in the whole startup ecosystem. Industrial transformation is like the big thing, right? Everybody thinks AI is the big thing, but industrial transformation is 100 times bigger because it's basically taking the Industrial Revolution and doing it all over again with a big giant upgrade. Would you put AI in the category of industrial transformation, or do you think it's inside and outside of that?

I think that AI is a key part of the industrial transformations that are occurring right now. I think it is also somewhat separate. In some ways, I think what AI does is it's fundamentally an optimization technique. I mean, that's really almost all AI comes down to. We're optimizing a process, a mathematical function, code to do something, whether it's predict the next word or figure out who has cancer. So we can apply those optimization techniques in industrial transformations, in optimizing factory workflows, in optimizing supply chains. You know, it's one of the big areas that we're now seeing AI being used is in optimizing logistics. It's been around for a little while and it just keeps becoming more and more important. On the other hand, one of the things that I think is important about AI is that it also can provide new capabilities. And a lot of industrial transformation, at least in my experiences with it, is focused on how do we more efficiently do what we've done before? One of the key changes with AI is that we start to expand our horizons of the possible, at least in the, in the good cases, so that we start to really understand that we might not just be able to replace a human with a machine. But rather get a human to do so much more than they could before by being augmented by the machine. And so I think that there's opportunities with AI to do much more than we're seeing in a lot of the kind of standard industrial transformation approaches, but it's also a really important element of successful transformation.

Interesting. I love that. So it's almost like when computers came out and we became 100,000 times more effective. And AI is like another boost, right?

Yeah. And, and I think, you know, the introduction of computers, or let's be even more granular, uh, the widespread use of spell check. You know, spell check means that we all lost a little bit of our spelling ability. I know I did. Uh, on the other hand, it means that we can focus on what really matters, which is the expression of our ideas. And we're not bogged down necessarily by the details of how we spell something. Word processing is similar. A lot of people who are typists lost their jobs because of word processing programs. On the other hand, it makes it much easier to rapidly iterate and revise. And what we've seen is large changes in the number of people who are able to express their ideas more clearly, faster, and to a broader audience. So I, you know, computers don't just let us do all the things we did before faster. They can lead to these real changes, almost phase changes in what we see as possible in business contexts.

That's really interesting. So let's, so let's move from there. So if you are an entrepreneur right now and you're looking at AI opportunities, the first thing you have to think about, right, is the considerations for building trustworthy systems, right, that align with the social values to your point earlier. So what are some of the key considerations an entrepreneur should have when looking at their responsibility for this powerful tool, you know, to prevent it from being weaponized, you know, as an example?

Yeah, so I think the first thing is, and I think this is probably good advice for any entrepreneur, is know your audience, know your customers, know your clients. And by that, I mean, know what matters to them, know what they're trying to achieve, and know how what you are building will actually serve their needs and interests, not what you think their needs and interests are. So there's a long history of people thinking that they know, oh, here's what our customers want, when in fact the customers wanted something very different. Now, of course, you can't ask people necessarily what they want. People aren't always good at expressing it. As Steve Jobs famously said, people will know what they want when we give it to them.

Yeah.

And I think that that's right. But that's all the more reason to engage closely with your customer base, with your potential clients, so that they can almost help co-design what it is that you're building. You can show them, hey, if we could do this, is that something you would like?

Do you think that AI could help with that? Because that's a big problem, right? Like, to your point, you know, it is almost impossible to get customers to tell you what they want. It's, You have to decipher and translate what they tell you to turn it into something that you can actually do. Do you think AI could help with that process, like conversationally or something?

So I think AI can help with that in a couple of different ways. The first is that you could use conversational generative AI systems to basically have free-flowing conversations with your customers and your clients without it requiring your time and energy, right? So you basically get a chatbot to elicit a lot of these things and to explore scenarios. So you could design, you could have a chatbot that essentially, you know, an LLM, for example, large language model that essentially generates a whole bunch of scenarios and says, in this case, what would you want to do? In this case, what would you want to do? The other place that I think AI can really help is in rapid prototyping and exploring the possibility space. So AI systems, especially generative ones, can very quickly create mockups of systems, can create sort of simulations of new capabilities. And so you can put people into a more realistic, a more naturalistic setting and say, you know, would you use this in your workflows rather than just asking people, hey, is this something you would like? Which I think can often suffer when people have failures of imagination.

They might not realize what kind of a cool thing they could So I wonder what regulations a founder should be cautious of, and how would they figure that out? You know, because they're passing regulation, right? So if you're building an AI system, is there somewhere you need to go to make sure you're not going to get yourself in trouble for building something that is, you know, using potentially the entire internet for its knowledge base?

Yeah, it's a great question. Uh, it is one that, uh, I have found myself spending far more time than I expected over the last couple of years thinking about and talking with folks in places like DC about what's going on. So the first thing that is really important for an entrepreneur to know is that right now we're seeing basically the development of a patchwork approach to AI regulation that depends a lot on where you're located. So the approach to regulation in the United States is very different from the approach in Germany, is very different from the approach in China, is very different from the approach in India, just to name 4 locations. One of the first things is to recognize that just because you've checked all the regulatory boxes to deploy a model in France does not mean that you're going to be on safe ground when it comes to the United States. And the reason I wanted to make sure to call this out is obviously the idea that different nations have different regulations is not anything new, but AI and AI companies are not bound by borders in the usual ways.

Yeah, yeah. That's a really good point, right? Because it's, it's like free.

Exactly. And it's very easy to build out an AI system and suddenly find people around the world using it. Especially if you just throw it up on one of the standard cloud services, it can be accessed anywhere in the world. And so I think that there is a much more complex regulatory environment that entrepreneurs are entering when it comes to AI than many of them realize. And so I think that there's a large amount of thought that entrepreneurs need to put into how are they going to make sure that they don't step on any regulatory landmines. Now, I mentioned the different approaches. Let me just briefly say, you know, for example, in the United States, the regulation of AI is happening in a very sector-specific way. So if I'm building an AI system in healthcare, I'm going to have to know about HIPAA, the health privacy law that we have in the United States. In contrast, if I'm building AI that's going to operate on social media, well, in the United States, social media is largely unregulated. And so there isn't going to be any particular regulation on my AI system. In contrast, in the EU, there's what's called the AI Act, which was just recently finalized. It's going to be going into effect in the next couple of years. And the EU AI Act is much more focused on AI writ large. It's a sector general kind of regulation that essentially says if you're going to be using AI systems in high-risk ways or in ways that create systemic risks. So if you're building a generative AI system that's going to be at the foundation of many different models, Um, then there's a whole set of auditing and testing that you have to do before you ever deploy it in the first place. So if I'm an AI developer in Germany, my brain is thinking about, all right, what are the high-risk use cases that this might be put to, even if I don't intend it to be used in that way? Whereas in the US, what I'm thinking is, okay, where are the sectors I want to deploy and what are the regulations that are specific to that sector? Now, you also ask, I think, a question which is very near and dear to my heart because it's part of how I got into thinking about issues around ethics and technology and ethics and AI in the first place, which is how should we think about the ways that our algorithms and models might end up being used for purposes that we never intended?

Yeah, that's what I'm worried about.

And you're right to worry about it. Um, that's, uh, part of what drew me into ethics of AI is in part learning that an algorithm I had helped to develop probably was used for purposes that I didn't necessarily agree with. So the first thing to say is that's less of a risk, I think, actually for many entrepreneurs than it is for folks like me in academia, because entrepreneurs are typically building products, which means that there's going to be a lot of work that goes into translating the algorithm and model into the particular use cases of your customers, your clients, whether those are the businesses Or consumers. In contrast, academics, we're much more interested in what is the general algorithm that could be used in many different contexts.

Oh, I see. Broad.

Exactly. Um, you know, AI is great because it's kind of domain agnostic, right? In general, data science, machine learning, we kind of don't care where the data come from. Uh, we're operating on the patterns that we find, and patterns are independent of what they're patterns over. At the same time, obviously you're going to tune your algorithm to the particular assumptions you can make, the particular features of the problem. And so as you get closer and closer to an actual product, I think the domain generality gets less and less. Even with the things that we think of now as the most general purpose AI systems called foundation models or generative AI, the large language models. Even those systems, we are increasingly seeing that very few actual applications in business or in society generally, very few of the applications are based on the untuned foundation models. Most of what's happening, even with something like GPT-4, is that companies are fine-tuning it in ways that make it much more focused, specific, and accurate For their particular needs. So I think we're, even though we talk a lot about general purpose AI, I don't think that it happens as much in practice. And so as long as the entrepreneur is retaining some measure of control over their intellectual property, they can help to shape where the technology goes.

Basically like commercial AI is relatively low risk, but then broad AI Is more high risk? Because one of the things that I notice is that if I aim the AI at data, it uses that data to return my answers. So if you aim the AI at the entire internet, we all know there's a lot of trash on the internet, right? Am I getting the truth? Or am I getting somebody who has, you know, carpet bombed the internet with the way that they think, right? So we saw in a couple of election cycles, we saw this, Massive insurgence of flooding the internet, you know, carpet bombing the internet with, with certain belief patterns that it seems like if AI was in place and somebody asked a question, that would be the answer that they would get because there's so much data on, you know, reflecting that answer. Like, how do you think about that? I mean, that seems like a pretty scary scenario and dividing, you know, Even more so.

Yeah, it's a massive problem. Just recently, Google came out with their AI summaries that's supposed to help with the search results. So it summarizes the leading responses that were found through search so that you might not even have to click through. And it's done things that have suggested, for example, that you should perhaps add glue to your cheese on your pizza so that it you know, is gooier and flows better, which, as I understand it, actually is true of advertising. Advertisers often do things to the food in order to make it look better. You don't want to eat it, though. That's a case where there's information on the internet that is being misused or misapplied, shall we say, by the search summarization system. Of course, it goes beyond just the sort of garbage that's out there on the internet in various places to these systems actually completely They're just trying to predict what the most probable next few words are supposed to be. So they will make up things that sound plausible. We call this hallucination. It's not the best term, but that's okay. They'll essentially make things up. Rather infamously, many of us in academia, when things like ChatGPT first came out, we would say, oh, what are the papers in my area that I should know about? It would generate a lot of very plausible sounding paper names. None of which existed. And in fact, recently one of the CEOs of the big tech companies came out and just said, we're never going to solve these problems. These systems, these generative AI systems, are always going to hallucinate. They're always going to return problematic output in at least some cases. So it's a huge problem, and it's one that we're never going to get rid of. Having said all of that, We can do so much better than we are right now. There are a lot of techniques and ways to help people know for these kinds of systems what we call provenance, sort of the history of where the information came from. And that's one of the big advances we've seen in the last 6 to 9 months is improved systems that can reveal the provenance of information to people that are using these AI systems. They can These systems can trace back the history of, well, the reason I've given you this information is that it came from this site. This site is reputable for these reasons. And we're starting to see the emergence of technologies that can help people more critically evaluate the information sources that are being put in front of them.

By like transparency, just giving us, because that's one of the things I worry about, right? Like if AI is doing the thinking and And we're not, are we just going to get dumber just like we don't spell as well?

So our cognition is going to change. I quite deliberately didn't use the term dumber. Um, you know, there's, there's a phenomenon that's referred to as deskilling. It's the idea that if you have a skill and then a machine or some automation starts to do that skill for you, or even another human starts doing that skill for you, that over time you'll lose that skill. You will Deskill. That happens all the time. When I was younger, much younger, um, I could drive a manual transmission car pretty well. I still can kind of drive one, but you wouldn't want to put me behind the wheel on a regular basis at the moment because I've deskilled. I've lost that particular ability. That one though, honestly, I'm okay with. Like, it doesn't bother me. I very rarely need to drive a manual transmission car, so I'm okay with deskilling in that.

Selective deskilling.

Selective deskilling, exactly.

Yeah.

So what that points to though is the importance that as we're bringing AI systems into our workflows, into our companies, into our daily lives, we actually need to be intentional and thoughtful about what skills are we likely to lose as a result, and are we okay with that? So the 3 most powerful words when we're thinking about technology adoption in my mind are, Are we sure? Just stopping and asking, are we sure that this is a good idea? Are we sure it's okay that we lose that skill? I'm really glad that people thought about that question when it came to the rollout of autopilots in airplanes. Autopilots can actually land airplanes better than most commercial pilots would be able to. But we don't allow commercial airlines to land on autopilot because we think it's really important that the pilots not de-skill. In terms of landing an airplane, like that's the one skill they need to make sure to never lose is to safely land an airplane. And I think that that's emblematic of a discussion and a debate that I worry is not happening right now. AI is getting rolled out all over the place. We are going to see some significant deskilling occurring. We're going to see some significant new capabilities arising because of the AI systems being used in widespread ways. And what we need to be having is an open, honest discussion of which skills is it okay to lose? Which capabilities do we want to make sure to gain? And have we struck that balance appropriately? Or are we just throwing the technology out there, seeing what sticks, chasing the latest bit of market share in ways that are ultimately going to come back and we're going to look up in 10 years and say, oh God, we don't have anybody who has those skills anymore. I mean, I think where I feel this most acutely right now is the rollout of code assistants. So systems that help people write computer code.

Yeah, I use those.

Um, they're great in some ways and they can really accelerate the writing of code. I've used them myself. On the other hand, I worry if we end up in a world where everybody de-skills in terms of writing code. I think that's problematic.

And then you get to the point where the code's writing the code and then you've got no transparency and nobody knows how to write code anymore.

Exactly.

All right, David, I want to get into one of your many areas of expertise, and that's time series models, but we'll have to dig into that next time. Also coming up in the next episode of the Startup Science Podcast, David Denks shares some valuable advice for founders.

So we really are at a place right now that if you want to be able to, to get ahead, the best advice I would give an entrepreneur is Figure out if you have access to data that nobody else has figured out how to use yet. The Startup Science Podcast with Gregory Shepard is brought to you by startupscience.io. Founders can log on to startupscience.io and use code Forbes to get free access. To connect and find out more about Greg, go to GregoryShepard.com. The Startup The Real Science Podcast is a production of Forbes Books.