Chatbots, Synthetic Users and AI for User Research with Hassan Aleem

Hi, this is Samuel and welcome to a special episode of the Behavioral Design Podcast. This episode is part of a new miniseries that will be dropped in throughout the season where we get the chance to ask our colleagues at Neon's Behavior for their unique perspectives. And as you know, Neon's is filled with some amazing leading experts and senior behavioral practitioners, so it's always exciting to hear from their perspectives and hear their insights.

And our first guest is someone I'm really, really thrilled to talk to, which is Hassan Alim, an incredibly respected behavioral practitioner with a PhD in neuroscience. And Hassan's experience spans years of research in neuroaesthetics, and it's also worked across diverse industries like fintech, health, wearables, and public health. Hassan's deep expertise in applied bioscience combined with his rich tech background, I think brings a really unique and valuable perspective to the table.

And this was honestly one of my favorite conversations we've had throughout the podcast. You can expect us to dive into some quite fascinating topics, including Hassan's thoughts on AI and its impact on the field as a whole, but specifically how AI can be used in user research and having made a chat bot himself with AI, what makes a good AI chat bot? And going deeper, what is the role of synthetic users and what is a is potential to do literature reviews, As you can

hear much fun stuff. And finally, we tackle a thought revoking question, but one that I would argue Hassan is probably best potential in the world to to answer, which is can I

understand and a sign for beauty in the world. So get ready for a thoughtful and I would say inspiring conversation with Hassan Alim. Let's jump in. Happens to Mugatroid. And I have to say, welcome us on to the Be Able to Sign podcast. Yeah, thank you. I'm excited to be making my debut. Yes, very exciting. And yeah, been looking forward to our conversation and I'm sure we're going to cover quite a bit

of things. And I guess it's it builds upon the conversation that we recently had with Amy Bucher on the podcast. And I was hoping to get into one of the topics right away. I brought up this idea of a lot of people in the behavioral science space right now are probably looking at AI and feel like, wow, that's interesting, that's cool. But kind of what is my role in

it? This is there a place, is there a role for behavioral science or behavioral scientists in the world increasingly with AI at the front and center? And so Amy made kind of a case for like, yes, it is very important. Actually, it's even more important in some ways. And I kind of wanted to hear your thoughts, maybe jumping straight in with a hard question. What do you think about that? What is What do you think about the role of Bible science in the

world of AI today? So I would agree with Amy and and you and Aline that there is definitely a role to be played in an essential role. The work that I've done has shown me is that like the technology's amazing and it can do a lot of the things that we might consider as behavioral design, but actually fall mostly under like what usability and

heuristics about, you know, like make it simple, make it intuitive, etcetera, all those things. But the broader behavioral design around motivation and engagement and barriers and all those things that does not fall under the type of capabilities that most of these tools have. And like what they're doing well, at least at this point, the thing that they AI does better than anything is make

things super, super easy. You know, so that is obviously a behavioral science design oriented thing that we want to make things as easy as possible, but it's not just a lot of ease. So I do feel like we're behavioral design and behavioral science has a lot to offer is of course, as Amy pointed out in like the domain expertise so. Yeah, yeah.

And I guess I want to unpack what you said there because for you, what do you mean when you say kind of that AI makes it easy or like they it's good at at that side? Yeah, I mean, so like an example is with ChatGPT and it's whisper mode and you know, it just makes it super easy to ask questions. It makes it conversational, which is we haven't had that ability before, right. So the ability to communicate in natural language, I think that removes a lot of barriers.

So it's, it's a little bit, I do think there's a lot of great things happening where like behavioral sciences, fundamental component, behavioral sense is built in into making things easier. But at the same time, like a, a, a tweet that I remember I shared with you that I saw it was about like when whisper mode first came out. And you know, whisper mode is this paid plan on, on ChatGPT

where it's just very intuitive. For example, I use it with my daughter and when I say, Hey, you're talking to my daughter, it changes its tone and talks in like a kind tone. So it's like a really, really well done. And so there's this tweet that was like showing the stock price of Duolingo and saying like, oh, Duolingo is cooked. Like it's over. People are just going to be able to talk to Chachi BT and just like learn languages.

And I thought that was funny because it's like, yes, of course, like this is fantastic and I use it all the time. But learning a language is probably 1 of like the most common behavioral goals. You could say like we have as as humans, like everybody wants to learn another language. And if it was that easy, you know, I, I think we would all be multilingual people, but clearly it's not an easy thing.

And looking at Duolingo, like probably the most gamified, most like behavioral design oriented company there is in terms of apps out there. And that's for a reason, right? Because it is hard and people need that motivation. So like, yes, it's super easy to just pull up my phone and start trying to talk in Spanish or whatever language I'm interested in learning. But like, there needs to be that structure. Yeah, no, I think that's, that's true. I think it's a great example.

I actually, I got my mom set up for advanced voice mode with GT for that reason, because it's been interesting to to see her journey with Duolingo. And then with advanced voice mode, I kind of wanted to set it up for her to see, you know, also for me as well to like as a mini user experiment to test, test and see what my mom would do with it a little bit, maybe as you would with your daughter and see how that interacts.

And it is interesting. But I do think to your point, it is incredible what is capable of, but it's still missing that kind of layer that supports long term consistency and engagement and supported learning in terms of a, some form of a learning plan and some form of a way to feel. Also, I think while the voice mode is generally, I think set up to be relatively rewarding in how it interacts, like it's

always encouraging and positive. And even if you spoke Spanish in the worst way possible, it will probably say something like, Oh, you're doing so well, good job, keep at it. And so it is rewarding in that sense. But I do think it, as we, as you mentioned with Duolingo, that Duolingo has done a good job of, of creating a, a kind of from start to finish, like just various ways to keep both short term and long term and motivation and rewards in place.

So even though the advanced voice one maybe delivers some form of like in the moment higher utility or some of higher value, it's kind of missing the the full delivery of long term behavior change in the same way as as the lingua does. Yeah, exactly. I, I like the way you put it in terms of like short term and, and long term. And, and also, you know, as we saw already, Duolingos adopted that you can now have like a conversational agent.

As soon as the technologies appear, the people that have domain expertise are going to adopt them and integrate them. So I think naturally in that sense, like when you have to specialize you, you have to build in domain expertise and and the behavioral science component. Yeah. And I do, I do think like it fits the narrative that we have right now where every week there's some form of like big headline saying this is the end of this news from Open AI is the end of this basically.

But I think bringing some new ones to the reason why some things maybe last while other suffer platform risk is I think that kind of Babel component, whether you know is, is the solution as somebody Duolingo, like is it actually supporting some form of meaningful behaviors change?

And on that topic, if I wanted to go from 1 chat bot to another and ask you a little bit about this really cool project that you've been involved with creating a chat bot for kind of a sustainability related projects. Would you mind sharing a little bit of that process of of developing a chat bot facing tool?

Absolutely. So I've been working for the about the past year and a half with sustainability NGO and we've been working on a chat bot to help answer people's questions about recycling and waste in general. And to me that's been a really interesting project because it falls under at surface level, this is that sort of AI gold rush, right? Where people are looking to see like, OK, well can we adopt these technologies and put them

in and integrate. But what I see the difference is like the approach that this organization has been taking has it's very different from what I would say just sticking it on. And you know, my role in in that has been trying to really put in that behavioral design and bring in the domain expertise as well. And I think that's super important because what might seem like the right answer, for example, like the example that Amy gave, the domain expertise there matters a lot.

And that matters a lot when it comes to questions about recycling and what to throw and what not to throw away. It's very, in some ways it's very much like a, a flow chart.

So you can't really have an LLM riffing too much there. But then also there's broader hierarchies of like how we want people to do certain behaviors, right? So, and there's like a segmentation that we've done where it's like, you know, the most devoted waste person is, you know, composting and doing all these things and like, you know, they might be motivated to learn a lot versus like who is at the very other end of the spectrum who just wants an

answer to a question, right? So we have to be able to design for all those things. And that's not a trivial thing to do in the chat bot because you it doesn't know immediately which one of those people is asking. So the goal is like how can we create like a environment of digital experiences where we can get people from that entry point to like more and more curiosity and more mindful consumption and does broader level goes? Yeah, that was really cool.

And I guess for those kind of part of the segment that you mentioned before about that are like almost maybe not don't even believe in climate change or something like they're very anti recycling. Maybe they can be forwarded to the debunk bots that we actually had. Gordon Pennycook talked about recently about supporting people in helping them change their beliefs around conspiracy theories. But I, I think that's really

cool. And I guess going back to what you said also in the beginning, there terms of some misconceptions people have about chat bots. I think as you said, like it's very easy for a lot of companies and organization to want to slap on a chat bot as they can have a quick fix, especially with the promise of AI and what we we can do today. What other misconceptions or mistakes have you seen people make when thinking about like chatbot as a specific kind of interface or a specific kind of

intervention? Yeah, I think I would have to go back to that same reality check of like, is this the best way to do it? And like thinking about this idea of like a form factor trap, like are we trapping ourselves into thinking about what is chat like? What is chat really right? Like moving away all of the hype, everything chat is talking to somebody and so is the best way to get at this to talk to somebody? Or is the best way to get at this to just click a button,

right? And I think relative facts like that, I think there's often times this over integration where you might say could we just have people click and through like 3 questions and we would know so much more than we would have to like make them sit through a conversation. I can't think of any specific examples off the top of my head, but I do think what can tend to happen is that shift towards like making everything a little bit too conversational.

And I do think that there is some probably unconsidered costs to like fatigue that might come with repeatedly engaging with no matter how natural something is. Like, you know, there is obviously a dimension of like we all like to talk to people more or less. But I also think there is probably going to be some emergence of like fatigue with

talking to chat bots, right? And we haven't probably hit that yet just because these as much of I think we're still all in a, in relatively a bubble of of like usage with these things. But that's the other thing as as well as they just being mindful of like are we just doing it because it's fun or are we doing it because it's the most efficient way? Like can we not get it in the way of the user actually by putting this technology in there?

Yeah, I guess like something comes to mind for me or two is like maybe one is the classic Clippy example where you're just trying to edit your document and then this whole like the little Clipper thing comes up and it's like you have to interact with it. The other one is where you call a lot of customer service numbers these days.

They are basically AI generated some form of like welcome process where you have to like, first you have to hear some AI voice say like, Hey, your call matters a lot to us. You're a valuable member of La, la, la, la, la, la. And you're like, I don't care about this. It's like, we care so much about you. And it's like, there's a lot of people calling right now. You have to wait 45 minutes and it's like, well, you, you said you care about me, like what is going on here?

And then you have to listen to even more stuff before you actually get to where you want to go. And you obviously don't want to have that user experience or customer experience where you used bombarded by fake interactions where you used to want to get some form of basic answer or some basic thing right.

Yeah, Yeah, I like that. And just to bring it home to the project that I'm the working on the sustainability recycling project, like one of the things that we're working on is this sort of ecosystem and one aspect of questions is object specific, right. So I might have a water bottle next to me and like, do I really need to talk to a chat bot to understand that? Or is there a faster, easier way where I just want a yes or no answer?

Because I want to just know immediately about this specific thing, right? There's some questions that I might be like, hey, tell me about plastics. Can they really be recycled? Or like, why can't I throw away a vacuum cleaner in the recycling bin? And I by the way, like these are real examples. Like 1 to a. Facility and people will throw away I expect their vacuum cleaner to be recycled. So like that's a different question, right.

So one of the things that we're working on is like this counterpart where like the chat is conversational, but then there's a QR code or link that takes you specifically about an

object. So when you scan that QR code from the object, it immediately tells you a yes or no answer about that and that's it. There's no expectation of like, hey, hello, how are you what's your name What's the it's like you just get the answer but at the same time like you want to give going back to like the behavioral design and curiosity you want to give the people opportunity to say like, wait a minute, I was expecting a no like why is it this way?

So I I do like the idea of like, you know it's not getting in the way, but it's there if you need it is is how I think about integration is nice, right and Clippy clearly like is getting in the way. It's literally undocumented. I remember every time like get out of the way, but this is Clippy's redemption art so I don't want to hate too much. Yeah, no, but I, I love that and and also think, yeah, this project is a great example of used.

I think I've had a little bit of obviously seeing a little bit of this and what you've been doing, and I think it's just such a good example of where having a behavioral science lens can add so much. You know, we think about it in terms of something like recycling as something relatively simple, but when you actually start to unpack it, no pun intended, then you really see like how many parts to this that is quite complicated in terms of cognitive aspects of

the decisions we make. Some of our understanding can be very different about similar objects or similar things. And how we view and understand all of these contextual components makes it quite tricky to design solutions for it. And I'm not even going to mention like the regulatory challenges in for example, in America, we have each state or each county might have different rules and so on. So, so yeah, so much admiration for this work and, and thanks

for sharing a bit of that. I guess to pivot to something a little bit different, I wanted to hear a little bit your thoughts about AI as a role in research. I guess from a perspective of applied aeroscientists. I think it's also important to talk about what can AI do for us to make our research easier or better. I guess you just said the same. I guess we're probably both you and I see AI as a friend of male scientists. Like would you agree? Yes, Yeah, I would.

Yeah, I agree. I think probably a friend of all sort of researchers or anybody who's thinking at a strategic level especially. But yeah, I think it's interesting. There's so much there. It's some, some things I am hesitant of, but some things that I am excited of and, and use. So it's just like, yeah, it's early days and you have to be careful in some things. But yeah, I I definitely see it as like a net positive as as a support. Nice. What about synthetic users?

Good or bad? Or like, how do you think about that? Yeah, my initial reaction was that. And since then I've been thinking about not more so like I I think probably getting past the sales type and, and which was like you don't have to do user research anymore, etcetera. But now I'm thinking of it more in terms of like for things that we have a lot of already

validation of in terms of in the training data, like it might be a good, just a reality check, right. When I run a survey, I don't just put it out. I, I have somebody take it live or walk through it so I can do some cognitive testing and understand like did they actually interpret the question as that intended? And so I think of synthetic users as another way to do something, sort of like, is this getting at the top type of questions that I'm asking, or am I just asking the wrong

questions? So I think yes, but you have to be careful about not using it as the final word, but more as like a validation tool. At least that's how I see it. Yeah, because I guess the like to impact this and for anyone who hasn't come across it is basically the idea that you take some form of, as you mentioned, some form of data around pre-existing knowledge about certain type of users or people

or from research. And you try to basically create a data model that helps with, yeah, testing various either product or research ideas with some form of instead of going to real people, real users, there's some form of like synthetic type of uses. And there's different models for this, some different ways that you saw some recent simulation model that just came out around this. And so there's a lot of work that is trying to simulate

people's behavior and simulate. Use your preferences and interests and personalities for some of the stuff. Yeah, to your point, I think I, I found it a little bit hard to immediately not feel like that this is terrible idea. I was like a little bit like you like the first initial thing was like, wait, are we just going to create some form of fake world and then tests what could happen in the real world, in this fake

world? And then instead of doing things in the real world, we're just going to take this fake world insights and pretend that's true. And especially when it comes to also known from bioscience that we live in this kind of weird social science world where a lot of research has been done predominantly on research participants that are college students in Western educated countries and so on. The risk is that data in data

out. So if you have like really biased and like weird data, you're going to get like this synthetic users is going to potentially act in the same way and it's going to further be reinforcing some stereotypes of reinforcing certain things that kind of is not really what the real world predicts. And especially, you know, with each context being quite unique as well, how much can we really

replicate and test? So immediately I was like, oh, this could create really bad reinforcement loops of like just not really learning from the real world, but just creating some more weird, like in all aspects, like weird ultimate reality that we're learning from. And then I was like, maybe there's some kind of steel man or some argument to be made around perspective taking and some form of pretesting of ideas, as you mentioned, where like it can be useful. And I've heard a lot of people

talking about this. I know Laura de Mollier, who I think we're going to have on the podcast soon, she talks about like, using AI for perspective taking and trying to get perspectives can be useful. But yeah, where are you leaning now? Like, would you use synthetic users in current user research that you've been doing? How do you think about it right now? Yeah, I think we can learn a lot from physicists, right?

Like in terms of this is in the same scientific art is like, OK, we understand some dimensions enough and now we're going to model and simulate and rather than going out and like a demolishing building, we're going to try to 1st simulate what that would look like and then see like where we should place the chart or like it's the same idea. But clearly like we do not understand human psychology to the level that we understand are physics laws.

So I think it's the yeah, the biggest concern, like you said, there is like I think the overconfidence that can come with being able to do something that easily and thinking like, I can take this away. And I think the the nice thing with other fields that do simulators, like they have medical medical models and they can say like these are assumptions we're making within these constraints.

Like we saw this, you can't do any much of that in with that level of control, at least in in this world. Then of course then like there's the weird aspect to it. That's just fundamentals. Yeah. Just not being able to simulate like half of the world if you were thinking about physics. So is it useful in that domain in that sense? It's not the scientific validity of it, like being able to take away real insights from.

I don't think so. But I'm still now cautiously optimistic about it. And I think with those rules about like thinking about what is it going to fall into in terms of is it going to be answering service or is it more going to be like a perspective taking and like a reality check or maybe like a cross validation thing, right where you can run a real study and see how close was I and what can I do?

And I'm hoping that in the future, like these companies get better and better as they have their toes dipped in both things where they're actually also running real studies and constantly updating their trainers. So I, I think there's hope there more than I initially thought. Very much had the similar reaction as you. I guess there's two considerations also here in

terms of one is the use case. Where I've seen some of the use cases now, it's been like simulating mobility related conundrums like how do people move in the LA morning traffic and what would happen if it was a closure on this road and how would that impact how people behave on some other highway where they can both now combine data of like, OK, what have

people been doing in the past? Like what historical data of traffic and so on combined with simulated agent stuff where they add some other stuff to it. Like in terms of we know that a lot of people have kids, like we know a lot of people have certain other things. So if we simulate people as more as agents as well with some autonomy where they might take some some autonomous choices, we might get a better sense of what actually would happen if this thing would happen or if that

thing would happen. So it would make for a better simulation somewhere. So I think that use case seems more potentially just straightforward that that could be very likely to be impactful. And the second thing, I guess is also considering like what is alternative?

Because if you look at some alternatives that research is doing right now in terms of either, let's say using M Turk or some form of like aggregator for research, that is like honestly quite low quality where OK, you can get insights, but it's like paying people sense to click things and test things and use getting some form of like really strange sample in some ways. So of of people that are doing some task for you that giving you some insights of human

behavior. But at the same time, it's however you constructed it tart fully obviously make that high quality by the constraints that something like Mturk would allow for. Or another side, you might consider something like focus groups, where, OK, focus groups might have their place. But at the same time, we also know that it's not probably the place where we will learn that

much about human behavior. Like people are quite influenced by each other and hearing people talking and all of these things like what do you get from focus group is probably relatively overrated from a behavioral scientist perspective at least. Conversely, if you do some really high quality, let's say user interviews or ethnographic research, can that fully be replaced by synthetic users? Probably not at same time. So I guess it all depends, as always, what you're combining or

or contrasting it with. Yeah, Yeah, I like that. And I think it really is going to come down to like who's doing it better, Like who has the best training data, who has the best validation, who's has the best domain expertise. And there is definitely, for example, the user interviews you mentioned, like in the same sense, not quite synthetic, but like there's now these tools and one company that I'd like outset

like where it's an AI moderator. And that's something initially I think any behavioral scientist or researcher might hear and like get a little worried about it. So I'm not going to let AI take over. But I think it's also one thing is like understand who's behind it. Two, it's about like the purpose, right? And the purpose that I've seen is that I can do myself, I can moderate interviews and I have a limit of how many people I can talk to before the quality goes

down. And it's just not the same, right? But it's also a very different goal like I'm going into in terms of deep inquiry for a problem that I understand really well for something like outside AI moderated interviews, like I could talk to, I don't know, 100 people and get results within an hour of 100 people's qualitative feedback for a concept that I maybe don't understand well

enough to go deep into. So like I think of it as like a pre deep inquiry step and and that's a sense it's a really nice thing. So similarly, I think for synthetic, I think it's just going to be about like not replacing, but augmenting your current process. And there's a lot of holes like quantitative qualitative research, like it's not perfect, like you said, focus groups.

So I think there are holes that can be filled with like these more synthetic or AI moderated research methods, but it's just about not replacing that supporting. Yeah, it's interesting when, you know, when it comes to user interviews, one of things I've seen over again by like really good user researchers is that they underestimate the power of their social influence on the the interviewee.

So if let's say I'm use a researcher representative of a company and you are the, let's say the interviewee in this context, and I say like, hey, I'm coming from this company, I would love to know, maybe we're doing some live user testing and you're interacting with my app or something. Tell me about like, how do you feel? Like what is the first impression of that? Like, how do you like it and so on.

What people are often blind to is that they don't think about how much they themselves influence the interviewee. In this case, you like you probably you could have the first impression to think this app looks shit, like really ugly. And at the same time you're like, but I don't want to make this nice person next to me feel bad. I don't want to like, criticize. They seem nice and so on. Like, OK, I'll you say, yeah, it's different. It's fun, it's cool.

And so you're not being fully honest with me because you want to spare my feelings, which is very nice human attributes, but that's honestly not really what I want to hear from a user testing point. Like I want to get your real honest opinion and I, we're really going to like understand what's going on through your mind when you're going to

interacting with my product. And so from that point of view as well, probably some AI mother is going to allow for more kind of candor or honesty because maybe at least people won't feel like they're hurting their feelings, you know? Yeah, I like that. I like that. Nice and maybe that's a final thing that's interesting is more in terms of called a literature review related collecting and interpreting knowledge for some form of projects.

Often times as a behavioral scientist, the first thing we do is making sense of, OK, what is the state of what the behavior we're trying to change, OK, we're trying to change this behavior or understand this behavior. OK, what can we find in the literature that has been looking at maybe this before? So we're doing some form of light or have a literature review. And yeah, I wanted to hear what you think about a is role in kind of making sense of existing literature for our products.

Yeah, I mean, I'm all for that and very upset that I didn't have this when I was doing my PhD at the speed. I think it's great for asking very direct questions. So especially when you don't know exactly how to frame a question, I think that's often the case with the real world behavior. It's easier to sort of like be in a very specific area of research and say like, what's the next logical question? You can use very technical terms, but it's like, you know, examples, what's people's

curiosity about recycling? Look, how do you even ask? How do you even do a literature search for that, right? So like just initial step of like natural language parsing, which I know you can set up now pipelines of like, how should I ask? And then you can put that in a research to like illicit. But so I think it's really good for getting specific answers as a gut check to what is out

there. And maybe this is just me, but I, I do think there is always, you have to maintain that balance of going into what these tools and recommended systems in general can offer you. And I know that you guys talked about in the in the previous episode and like doing a little

bit of your own work as well. And I think with science, often times good ideas come from reading tangential things and doing that exploratory aspect and realizing, oh, I would have never thought this was relevant, but it is leading in a direction that you, you connect the dots. So I think again, I think with the theme I guess I'm seeing is like not using as the final and the only approach, but using it as like 1 aspect. And what it comes down to is stepping back and defining what

your process is to say. Like, well, am I just exploiting and just asking specific questions or am I doing enough like open-ended exploring and trying to like cast a wide net as well? But in general though, I think there's no, I don't see any harm there because it's not like there's hallucinations, especially with these specific tools that giving you actual research papers.

So I think it's probably one of the best sort of like academic research use cases we have with AI so far, the probably the most advanced in terms of maturity. Yeah, I agree.

I think it's definitely a little bit of a fine line when it comes to it's great to maybe get AI to help us do some things, maybe help us to think better about certain things, but not having AI think for us. And I think that's where like we don't want to get crossing a line where we're used leaving AI to think for us and make all of the decisions or conclusions without any kind of critical thought ourselves. And that's what I think it helps to have.

I think to the point of you actually having completely appreciate, like I think you're benefiting from having a critical way of evaluating research and understanding research so that you can when there is maybe not a

hallucination. But I think I was to say most of the time these days they don't hallucinate, but they just maybe oversimplify or they misconstrue certain like things in a oversimplified way that you can like spot that and be like, well, OK, that seems a little bit oversimplified. So I think having some expertise, I think it's still a really, really, really big leverage point.

Yeah. And talking about your PhD and also about expertise, I guess it's a final topic before we wrap up. I wanted to touch upon, I think a very unique expertise that you have, which is with neuro aesthetics, a word that I think a lot of people maybe haven't heard of. I know that when we had some form of a LinkedIn bot roasting our our LinkedIn profiles, that it's basically said that Hassan has found a scientific way to say I have fancy thoughts about pretty things.

And this is when we we can ask people as a final question, kind of what's your most controversial opinion about AI? But I kind of want to actually frame it a little bit more for you because because on this topic, I want to ask maybe controversial question, which is, can AI understand beauty and can AI recreate beauty? Yeah. That's such a great question and definitely controversial. I, you know, a lot of the work that I started off doing was

about, was simulation, right? So we talked about simulation a little bit and we were simulating people's internal States and the environments that they would be in and then simulating the what they would think about certain images coming in, right? And then there's been really great models of like, what does it even mean in terms of beauty?

I do think it's possible, you know, if I'm not going to get into the quality of it, like are they actually experiencing beauty? That's probably, you know, I don't know how that happens in a silicon, but I think, I think I do think that they will, if you base it off the cognitive and neuroscience frameworks that we understand about what goes into beauty judgments. I think it will get to the

point. And I, I hope that it does get to the point where we have a agents that can appreciate beauty because I do think that's a fundamental component of humanness. So I do think that's a goal that

we should be striving for. And especially when it comes to like personal relationships, Like I would want to have an AI that could understand what's beautiful to me and like have its own understanding of what's beautiful to it, because I think that would give it much more personality than in an AI that like just didn't have an understanding of beauty to begin with at all. So I think it's a worthwhile goal. And I do think it's a possible

goal. It might be controversial, but I think I'm excited for, you know, sort of setting the parameters in place and see what AIS start finding beautiful and and see what their justifications for that are. I was really interested in, you know, all of the image creators and how they are deeply aesthetic. You know, that's sort of a funny bias that they have from our training. Like, you know, they just cannot make non good looking people and non good looking scenes at this

point, right? That's also this understanding is like their beauty is so unidimensional, like whereas humans we find beauty and ugly things and we find beauty and things that just would be totally unexpected because it's so individual. So that's what I'm really interested in is like a is that have this really like unique idea of beauty that's not traditional, that's not weird and just not like reinforcing stereotypes. I hope for that future. Yeah, and I echo your your

sentiment. And I part of me wants to have AIS with a little more agency and opinions in some ways. Like I feel like they're a little bit too agreeable as well today. Like if you ask AI anything right now, any type of AI, especially any AI shut but and so on, they will basically say, yeah, that's great, you're smart, you're fantastic and so on. And it'll be fun to take a photo of your living room and actually will be saying like, no, that's ugly.

That's an ugly decor. Like you should not have that pillow that makes your couch look terrible or something like that. So having some form of aesthetic and knowledge and preferences and, and understanding of beauty would be, would make it more interesting, fun.

And I guess a light version that is actually being used for today already is that there are these AI tools for where I think especially being just the one I'm thinking about, especially for men, where they can basically get photos of themselves and it will tell them like, Hey, how here's how you become more attractive. And in some ways, it's mostly saying like here, hey, you're not cutting your hair like you should probably cut your hair in in some way.

Here's a good look for you based on your facial shape, you should have probably had this kind of haircut. And secondly, you seem to have a very, very bad skin. Here's the skin care routine that you can follow to improve your acne or whatever it is. And so it's not giving you a filter. It's not giving you any kind of like this kind of dystorpium thing. It's just giving you a little bit of like notch be like, Hey, cut your hair, trim your beard and take care of your skin

basically. And here's some advice based on your facial shape and and kind of bone structure of your face that would probably suit you quite well. And that is early kind of sense of like giving some form of aesthetic recommendations, which I think is interesting. But again, like what I think is dystopian and, and scary is where it becomes too narrow and, and too homogeneous and it just becomes that it gives the same recognition for everyone.

And so we're just living in the society where everyone has the same kind of aesthetic look. But to be fair, we're already, I don't know, IKEA furniture is in most people's homes, sadly. And even as a suite, I think that's a little bit sad that everyone has the same kind of furniture and and same things in various ways. I think maybe again, we're coming back to that. In reality, we are quite

homogeneous in some ways. And yeah, it is interesting to see how AI is going to shape either that becoming more so or maybe going the other way and and supporting us to have more unique tastes and preferences based on our unique situations and contexts. So yeah, we'll see. But hey, this is so much fun as I could speak to you for a long, long time and I'm happy that I have the benefit of talking to

you on a weekly basis anyway. So I'm sure we'll follow up on this conversation, but for now, really appreciate all you do in your work in the field and for coming on today and sharing your fantastic perspective and expertise. So, so yeah, thanks. Yeah. Thank you so much for having me. Thank you. Quick word on Nuance Behavior where we help organizations build impactful digital products using behavioral design.

We only take on a few clients at a time to ensure the highest level of quality for our tailored evidence based solutions. If you'd like to become one of our special projects, e-mail us at hello@nuancebehavior.com or we could call directly on our website, nuancebehavior.com. Happens to. Mugatroid. Oh. Well, I'm very happy to say you

see, you see, you see, it's harder. It's harder than it looks. OK, OK.