 What's up, everyone? Welcome to Simulation. I'm your host, Alan Saakyan. Very pumped to be talking about social science and AI. We have Sonia Schmergerlund there joining us on the show. Hi, Sonia. Hi, Alan. Thanks for coming on the program. Thank you. I'm so excited for this. It took a long time. A year ago, you were teaching me a lot about your work and I was so pumped. It took a little bit of going back and forth. You were actually the catalyst that got us to the American Anthropological Association as well with your focus on studying humans and behavior. And then the last two years now we've been doing partnership interviews with them at the annual meeting and they've been blowing our minds. That's awesome. You have been there more often than I even though I'm an anthropologist. Listen to Sonia's background. She is a senior research scientist at Smart Information Flow Technology, SIFT, where she is a PI on myriad projects at the intersection of high-risk tech innovation and psychological well-being. Winning government research grants from NASA, DARPA, and the Department of Defense. And you can find the links in the bio below to her SIFT.NET profile page, also highseas.org, and also her LinkedIn and Twitter pages. Okay. So, Sonia, let's start by talking about kind of the main research areas and we'll break these down as we go. The first one is team dynamics and social sensing of human and human cyber teams under high stress. Yes. So, again, this is like a very macro-level perspective. Walk us through this field of study and the projects within it. Sure. So, keep in mind that I'm a social anthropologist, as you have noted in the beginning here, not necessarily a psychologist or a team researcher. I'm also not a computer scientist, but I have been working with a tech company now for 10 years and I'm trying to merge social science knowledge with, you know, next waves of development within the AI framework very broadly. So, very often what I do or what we do as a team is trying to see where the social sciences can actually inform new technology development. So, all of these projects have ultimately the goal in order to be applied, an order goal to have a tool that can be developed, a tool that can be used, an AI that can be used and implemented. So, what I do very often is like in early stages, come up with some crazy ideas as to what we should be looking at or what we could be looking at in order to inform the development of the technology, but then also see it through and then test it out in high-risk experimentations such as how I see it, for example, to get data in order to better understand how we work as a team, for example, but how we work as an individual as well. Because, again, as an anthropologist, I'm largely driven by, personally driven by understanding what I don't understand. Yes. And I quite enjoyed it. So, the Metula project, for example, is one such research project that I like to, and somebody asks me about this, I like to say I perform PsiOps operation in a secret research facility in Hawaii. And it's not even wrong, even though my funding agents probably don't want to hear that. No, what I do is I work as an API on so-called space simulation studies. We are trying to simulate the conditions of space flight, either to Mars or to lunar, for example, to the moon, lunar surface operations, and then have the stressor be the environment and the confined, the physical environment, the social isolation, have that actually be the stressor and understand how people are interacting. Now, that creates in itself already a lot of stress, and it gives us as researchers the opportunity to actually also have control over the things that we otherwise might not have control over in an environment. So, it makes it what we like to call ecologically a bit more valid. When we have subjects come into a lab, there's a lot of things that can go on outside your life that might impact the results that we see in a lab experiment, for example. It's also not very, of course, basic research is very important, but it's very different from looking at the desktop with specific images and then record a physiological signal based on an angry face that you have seen. It's not that this research is not important, it's very fundamental as well in order for us to make assessments about the human mind, if you want. It's a full-scale space simulation where you simulate literally the restrictive environment and social setting where I'm only going to be around five other people for months, if not years at a time, and the individual psychometric profiles of all these people have to dance. They have to harmonize. This is a good analogy, like music, for example, or dance, right? Like the ensemble that plays together, stays together. There's a rhythmicity that I believe we can pick up and we pick it up in language when we look at linguistic synchronization, for example. We look at the physiological signals that synchronize. We look at microbiome signals as well. Even the microbiome synchronizes. That might just simply be because everybody's eating the same diet, but there's also even studies that have shown that the air picks up signals that can be informative of social states. Philharmones, for example, is one such thing. I'm not the expert in that field, but I've seen interesting studies that have happened in movie theaters where people are watching a movie and experience a variety of emotional states and people that know something about toxicity in the air are able to pick up the signal in the air as to what the audience was feeling, in particular things like suspense or happiness, and that was perfectly synced. And spiritual people sometimes call this the collective consciousness. Yeah. There's a lot of interest in collective consciousness or collective experiences altogether. It's all a matter of scale. If you have a team of six people in an enclosed environment, that's a collective already. If you have a cultural group, that's also a collective. It is a matter of scale, but I think there's clearly something very beautiful about this experience, and it can be very beautiful, or it can be very stressful. A family is another unit. So where you live in close social relationships, where you live with people that have an impact on you or that you care about, Barbara B. Suxby at UCSD has said that social relationships are our greatest stressor, but social relationships are also our greatest release from that stress. And I think that is exactly one of the things that we are trying to understand when we look at collective allostatic load. Yes, yes. This is a very important subject, the allostatic load, allostasis. Teach us about what that is, and teach us about methods, which we've talked about on the program so much, methods to do behavior changes that return us faster to healthier, happier states of being, more pleasant states of being, instead of excessively ruminating on anger or stresses. And the importance of that, especially in these enclosed collective environments. Yeah, I don't know how I can teach people to become happier. That's not the business I'm in. I'm still trying to figure it out for myself. That said, I don't even mind if people have social conflicts, because then I can study that. You can study them. I'm not afraid of that either. But this is so crucial, because how do you create a, the proper people going to these lunar surface missions or Mars missions, the space missions that have the greatest efficacy for success. And how do we make the, not only the biometric psychometrics for that, but also the, build out the environment, the different things that they can play with, where I think you were teaching me about a year ago, that there's literally like simulations of like beaches and, you know, VR of experiences in that. So that, yeah, yeah, yeah. That was one of the studies that we have been carried, carrying out previously where we have looked VR, we've used VR, we are still using VR, then as a tool in order to combat the feeling of isolation, if you're just perfectly suited, if you live in a very enclosed environment, then you have the opportunity to, you know, escape. If, you know, the people that you live with start to annoy you, just basically literally, you know, don't just close your eyes and go off into La La Land, but actually go into a virtual reality space and, you know, feel like there's, you know, there's something else going on or you can distract yourself from like maybe things that are going on. Let me back up to what you were asking about these teams. So in the past, very often, you know, clearly even spaceflight and VR, all these things that they look, you know, they sound cool and it's all cutting edge, of course. But we are very often focused on, you know, the skillset that these people have, that they are physically fit and these are very, very important things, of course, because if we think about spaceflight, we cannot, you know, we have to consider these factors that they are mentally stable and that they are physically stable. But what we have neglected very often in this kind of research is the social variable. How well somebody is able to deal, for example, with social conflict has been like, you know, we've learned the hardware that that actually can make or break emission. And these experiences that we have and we all have had, like when working in a team, like we said before, for example, if you work in a musical example or if you play like Dungeons and Dragons with somebody else online, you can have a great experience with a small team or you can have like an also great experience. This is utterly important when you like send people. Like when you play the game of Monopoly and then the person that's like losing flips the board game over versus ones that can harmonize better with the process. Yeah, yeah, yeah. Yeah, it's very simple. Or like when you have a deeper experience of someone's like emotional state where do you maybe get triggered by somebody else's move or mistake in some way in these small enclosed environments? Or do you know how to stay emotionally resilient, meditative, equanimous, help them also grow and ascend rather than take them down into these stressful situations? Yeah, all very good points, of course. I mean, we are trying to figure out how the team can help each other because there's of course like stringent screening criteria. We're trying to get the best of the best people, not just for the simulation studies, but especially when we're thinking about like real space exploration and not just like individuals living together in a habitat in Hawaii, for example. Yeah. So of course we look at resiliency. But the thing is that we have very little data or very little knowledge as to what is actually happening to a person over time, over long periods of time, right? I mean, I can be resilient or at least try to do the best I can on the short periods of time. When we go to work, for example, and you have like a bad work environment and have a team that doesn't work well, you get to go home and decompress. But if you are continuously exposed to a stressful environment such as in like isolation in a confined environment, you don't, these kind of like mechanisms very often break down because you don't have the ability to withdraw and maybe recharge. And that is what the allostatic load really is. Like it's a chronic stress that you are exposed to. And over long periods of time, this is going to change how you react. And people very often themselves don't even know how they might react like when they are like locked up for 12 months, for example, which we also have had in the past. Now the next set of studies is going to be a bit shorter, but we are going to do one month long studies because we have a different like research agenda here. But ultimately, if we think about Mars mission, we are talking about two and a half years. And so far, we have very little data about what that does to the team, what it does to an individual living together for two and a half years because, you know, we haven't done that very often. Not at all. So the longest was actually one year. It was one year long. And this is also really mind blowing because you do things like you have simulated the environment to such great depths of nuance that there's even those initial, you know, they're wearing, when they go out, they have to wear the equipment that they would be when they were on the surface. They're wearing the large amounts of equipment. They have to go through the airtight chamber and, you know, the dirty off and then go into the oxygenated chamber. I mean, this is very, there's that type of stuff. But then there's also, you know, like you're mentioning today, we do things like walk when I'm in a feeling of maybe this stress, allostatic load is building up. We do these things called like blowing off steam where we go exercise or we go out into nature. We go on a walk. We just, you know, kind of like move apart from the other person for a period of time. But when you're in the small habitat and there's, you have to like go out to like Mars to like blow off steam. How, you know, there's, so there's that. Like what are you going to do in those situations? Are you going to go into the virtual world? How are you going to work out in a small little chamber, et cetera? And how do you also predict the right people in terms of like, we were talking about this last time, but just, if you're sending six people, do you send people, do you send six women? Do you send six men? Do you send three women and three men? But how, like, do you want them to have romantic relationships because then they can procreate on Mars? No. We don't. You're like, I know this answer. Why not? I mean, I mean, imagine the trip and I mean, it's just a safety risk. I mean, you cannot give birth in a spaceship. There's too many like factors. What about birth on Mars though? There's no infrastructure. There's nothing in place. I don't see unless we have a colony. Correct. Yes. Sending a couple trips with robots to assemble the initial, correct. Yeah, infrastructures and habitats. I haven't thought that far. So far, we are just trying to have people actually live together for a year. Survive a trip and back. And I mean, it's a controversial topic. I like to talk about it. Some of the funding agencies don't like to talk about things like sex, for example, because it's not public relations. But I mean, everybody knows about that. We are also sexual human beings with certain needs and that need to be addressed. So, you know, it should be sent mixed teams, just women, just men. I mean, the needs, the human needs are not going to go away depending on the team composition. Some of the simulation studies have shown that like romance can, you know, improve mood. Of course, that makes sense. However, even like as all social relationships, romantic relationships also undergo a life cycle and in two and a half years that could have catastrophic outcomes if you go past the honeymoon and then like the conflict that otherwise might have been minor becomes much more severe, for example. So the general tendency so far is to avoid that and have, you know, people that can compartmentalize that for themselves and don't engage necessarily in romantic relationships or sexual relationships. I'm not the expert in the field. I know that there's some other researchers that look into that a little bit more under cover again because like it's a little controversial as a topic, but I think it needs to be talked about. And, you know, I'm the first person who likes to talk about these kind of things, but I don't have much to say. Besides that, with regards to team composition, that we don't have a great understanding yet. The general tendency is to have mixed teams and that's not just from knowledge from the mass or, you know, space simulation studies but organizational and behavioral studies. That, you know, if you compare it to other data that we have and there's very little data on just all female teams, for example, because in the past if you think about space simulation studies it has been all male teams. Now we have realized that women are maybe socially a bit more intelligent so the team benefits generally with more social intelligence. Is it all female team? Would that be ultimately better? So far the verdict is out. I don't know. There's a lot of things that would speak for that but there's also a lot of things that would speak against that. You know, there's gender differences when it comes to, for example, you know, depression or mental, like mental things, I should say. And depression in order to answer the other part of your question, you know, how do you, how can we help this team? There's, it has been reported and we have experienced that like over long periods of time there is certainly a high risk of more than one individual maybe the whole team to start to feel depressed or what we would consider like, you know, low mood or low valence and what we can make assessments of. There have been some anecdotal evidence for that which I think is quite beautiful comes also from a study where there has, there have been like six individuals leaving together and one of them started to withdraw more and more and there's private quarters in some of these like isolation chambers where you can withdraw and this one person like was withdrawing more and more and like severely feeling depressed and then the team members of that crew had to basically intervene and they came together and literally lifted that person physically out of their quarters and said, come, you are part of us. We hold you here. And I think that speaks for the collectiveness of small teams of like us as humans that like we have these capabilities of saying like, you know, you don't feel good. It's understandable you want to be left alone but the more you are not part of us, the worse you might feel. So here we are, like we are able to carry you, you know, and we offload some of this allostatic load that you experience much more than we do. So going back to the research goals with the current project, for example, is exactly that like this offloading and unloading of stress. If there is a person that is depressed that is part of your team, is that person going to ruin it for everybody else? Or is there a mechanism in place where everybody else can like actually, you know, help that person in order to make sure that everybody is going to be okay? So you're also designing interventions to decrease allostatic loads, to return people to allostasis? Well, right now we just want to more or less understand the mechanism and once we understand it, we can then provide maybe diagnosis and then once we can provide a diagnosis then we can intervene. It's, you know, it's not as simple as that that we say try to do this. So, no. Yeah, yeah. For me it's been now several years of interviewing different people that have come onto the program that have been teaching methodologies for returning themselves to a state of equanimity or inner peace given this serious just swings in different directions that we experience in life and to hold that sort of inner state of deep love and peace and interconnectedness presence that is just the most difficult, almost part of knowing yourself is knowing how to get back to that state because then what happens is whether you're with family, friends, coworkers, people online doesn't matter your ability to be come back to that moment butterfly effects out to other people and then it creates that more harmonic environment around you and around the world by doing so. So, this is actually to know these mechanisms again, thousands of years have went with people studying meditation and this is, and thousands of years also with shamanic experiences with ethnobotanists with plant medicines we do more and more archeology into this and we understand that modern health care is missing this component. And so, again, it's one of these things that what can we really leverage in biology in these rich cultural experiences that can make it so that we can thrive better as we go on to these space expeditions and we have these enclosed environments this is huge for the topics that we've had on the program. Oh, absolutely. And if I take off my science hat but put on my Sonya hat, like, you know, I personally very much believe that the intention setting that we do when we meditate or when people pray or engage in some sort of like what we consider spiritual experiences, I think is, I believe, important. So, even if we look at, like, crew members and, you know, a spaceship the intention setting is important. What are you, why are you in it and what are you in it for? Is it for, you know, self-promotion? Is it for, you know, the good of the world if you want? I mean, these are things that might not be that clear for yourself even because it requires a lot of self-knowledge in order to, you know, know things like that. I certainly don't know anything. Yeah, yeah, this is... I'm really happy that we wandered into this field because as we design better and better mechanisms for this we can ensure an architect a more prosperous world. Okay, that's a lot on high seas and so people, you guys can find more information about that it's high, H-I-C-S-E-A-S.org if you want to find out more information about that project. Okay, we still have a lot more to talk about. So that was Medulla. I should maybe just mention that this website is still a little outdated because it has the high seas habitat is currently actually owned by the Blue Planet Foundation which is a collaborator on this project together with actually the behavioral science lab at NASA as well. So we are conducting like a number of studies like next year at the high seas habitat that is not necessarily associated with the University of Hawaii anymore. So the website is actually reflective more of what we have done in the past but as we go forward it's Blue Planet Foundation and also the Moon Base Alliance which is owned by the... Blue Planet Foundation and Moon Base Alliance. Which is who actually owns that habitat. Okay, got it. And with it like Hank Rogers who is a prominent figure that actually owns the habitat but he does a lot of work with regards to helping us understand sustainability not just in the context of spaceflight but also how can we test out these technologies of reusing everything not leaving any waste behind that is important for spaceflight but ultimately also for the planet we live on right now. Yeah, actually when you mentioned Hank to me last time as well I was hoping that we could do something with featuring people like Hank in this space as well. I look forward to... Oh you should, he's awesome. Yeah, these types of interviews are really crucial like what we're having right now. They're really important because they synthesize a lot of the most pressing fields plus the direction that we're going in the future and to know how to best be efficient in that process and reduce the oops moments. Okay, let's do another one in the same first point that we were in. Again, this is under team dynamics and social sensing of human and human cyber teams under high stress. So the next one is Vanessa. Virtual analysis networks and explanations for social sensing analysis. Okay, so walk us through this. There's not that much to say about the project. It is a smaller project actually and it's very related to what I've been talking about before. It's also related to a new effort called ASSIST. So the goal here is really try to understand not just human teams but human cyber teams meaning that if we have a cyber agent or an AI as part of the team like how does it change the team dynamics? What would that function be? So some of the research efforts are often going towards getting a better understanding of that where an AI agent would be more or less an assistant or a helper for a team and where we could like offload all the tasks to a cyber agent or an AI that the team itself is not good at such as gathering a lot of information identifying patterns, making inferences about team dynamics themselves and then provide recommendations what you were hinting at before with regards to what kind of strategies should we provide. That is more part of that project is like to do and I should maybe just talk briefly about some of these sensors that we are looking at. This has all been born out of really NLP natural language processing methods originally where we had teams or individuals write journal entries and then we analyzed the language that they used in order to make assessments about how they feel that then has grown into more behavioral research where we also looked not just at what is said or written but also what kind of behaviors we are displaying and now we are integrating also physiological components such as stress measures or as mentioned before even like the microbiome so we are trying to combine physiological cognitive and behavioral components ultimately in an algorithm that then can monitor teams that then can help inform an AI if you want about how the team is feeling in our case with a big focus on the social components and not just the health, the physiological components but on the social components of the team and then like provide recommendations or intervene before it goes down. That is the ultimate goal and that is what we are using these studies and these metrics from. Vanessa is really more the development of the AI using some of the data and so is actually assessed and like this is a new program where on the one hand like we are trying to build a model of like collective allostatic load with assist we are trying to build a model of like what is generally considered theory of mind and in our, I should say in SIFTS viewpoint it's a plan recognition problem and plan recognition is a concept coming out of computer sciences and like give your mind that everybody at the company that I work at is a computer scientist so I'm the only social scientist there which is for me if I translate it to you know my social science world is a problem of knowledge sharing so my world view if you want is informed by the knowledge that I have and your world view isn't informed by your knowledge that you have and I have my perspective and you have your perspective and we are trying to understand each other we are doing a pretty good job at that right now but I don't know I don't know you well enough in order to really understand what's going on in your head nor can I make an assessment about what you're gonna ask me next who knows what you're gonna ask me next If you had to throw a number out it's probably I know like 0.01% about your world view you know something like that then that makes it so hard to have these and it also makes it hard to have that stream of biometrics, psychometrics being analyzed by an AI to make recommendations about what to best do to decrease our static loads or to increase efficiencies towards goals yeah and you're pointing at the very important problem that we have in that space and that is the quantification problem it's relatively easy for us to measure a hard pit like if you are off the scope of it Alan listens to his own pit guys I've made so many posts about that we have a stethoscope here on onsite that we listen to heartbeats at night or each other's heartbeats and it's actually a really important exercise hopefully more people around the world will tune in oh you can also listen to your gut and you'll hear some very interesting things I would be really scared actually who knows what's going on there so there's the quantification problem like there's certain things that are easier to quantify now if you think of a hard pit hard rate variability but then like you merge into language and then it starts to become a bit more difficult to quantify right like we can say like okay positive negative if we talk about valence but if we start to try to understand intentions you know we can we're trying actually to do that in language as well it becomes much more difficult now if we remove language altogether and I have to make an assessment about like what you were thinking about just based on what I see and maybe your body language or you know your head nod or the size of the pupils of your eye and we are actually very good at that as humans because we have evolved to pay attention to these things but how we don't know enough about what we are actually paying attention to so it's very hard to inform or provide the quantification of these cues to an AI but that's the intention of like this one program in particular is try to use a lot of sensor data in a Minecraft play actually in a scenario in order to build and test various theories and models that are related to what are considered so-called emergent states where you know like situational awareness or like your merges of mental models your mental model my mental model and then bringing it back again to my own field with an anthropology I think this is very important because here we are talking again of different scales right like you and I versus you know China and the US for example these are also you know different perceptions that we are trying to make assessments about on a very large scale with very large variability within but that's a different set of you know collective consciousness if you want that is hard to quantify right because traditionally within anthropology we haven't been quantifying anything it's all qualitative you know and great records and I shouldn't like you know I want to point out I believe that it is important than ever to have this qualitative angle at that because you know there's a whole area of issues if we try to reduce complex knowledge and this is something we do a lot when building a SU you know yeah now I'm drifting on that's great that's like you said pointing at that specific putting the flashlight on reduce complex knowledge into some sort of a recommendation to augment experiences good luck I mean this is a very very difficult one of the most important fields that currently has a big spotlight on it structuring unstructured data making sense of it and applying it into real life to augment our experiences damn that one's a huge one okay so that was on Vanessa and on assist okay so let's do the second area machine learning bias and the influence of sociocultural context inherent in big data that is actually a good transition point with regards to what we're just merging into the field of anthropology you know can the idea here is that you know the bias in the data is a huge problem generally when building AI we know all about like the difficulties of like training sets reproducing you know racism or gender discrimination because the data sets that we have which are used when we talk about AI or building AI or recommendation systems or what not they are full of bias so the idea here is that can we advance anthropological research if you want if we don't look at the bias it's a negative but as a positive in order to tell us something about the society or culture it actually came from and I think it's fundamentally at the right approach because a lot of efforts actually by the big tech companies for example have focused on trying to remove bias in order to really work I consider it a little bit makeup in order to you know make them look better or in order to improve specific outcomes of say a search result for example that happened to be racist but it doesn't remove the structural bias inherent in the data because the bias itself comes from a social that we cannot change as easily so why not go and try to understand the cultural context it actually came from and what are the variables that contribute to the existence of the bias to begin with in this case for example gender bias because gender is a topic dear to my heart and I'm trying to understand so we looked at gender bias in like large amounts of Twitter data from all over the world literally 100 different countries oh you went to twitter for the data he can say a lot of things about that so what I should say is this is at end of by now a collaboration with Stanford University and the computer science department of Stanford University and they provided us with that data in order to analyze that so we had very large amounts of data and what we what we looked at is like more or less a lot many different things but the idea or ultimate goal is and was with the project is can we can we find some sort of like causal relationship ultimately or you know it's very very hard and I don't know if we really can or at least explanations that can help us understand differences in like how gender discrimination is quantified in the language that we are analyzing could we could we maybe say that like a root or an upstream variable in gender bias and discrimination maybe a if in the case of a man maybe their relationship with their mother things like that could be some of the most upstream or root issues if there was like a loving and tight relationship between a son and their mother growing up that man will likely respect women with great love that is a very good point and it's not one that we looked at yes yes I'm always going to upstream or root things but yes tell us about what was like that no I see you know in a different project we looked at in cell data for example you know the in cell community yeah in voluntary celibacy yes which I really hope can find a a that our society can find love to help the more healthy states of consciousness and being for people that are labeling themselves this way because we can do better I know we can do better yeah because I feel I feel a lot and I'm very sensitive to things like that and we can kind of dig ourselves into holes by associating ourselves with labels rather than kind of like taking ourselves out into challenging context and learning how to grow through hardships yeah yeah yeah I'm glad you feel a lot I feel a lot too but it's mostly anger especially when it comes to the in cells I'm a woman so I'm biased that said my point was you talking about like you know mother some relationships I think that is actually something that does play a role in that kind of community the data seems to indicate that that much I will say about you know things like there's your of course your upbringing and if you have been taught to respect a woman or not plays a role I mean no doubt about that but when looking at you know on a global level at like gender stereotypes and like how much man versus women associate themselves with particular gender roles that's a little bit different because I believe that gender is a social construct I believe that there's absolutely a biological reality that you know I look different in you and my body looks different in you but the way we behave or the way you know what interests us that is very often largely influenced by how we brought up and like we brought up to fall into this gender stereotypes and that in that sense looking at the global scale is actually very interesting because we have like direct comparisons between countries where the associations of the individual with a specific gender stereotype is maybe a bit more reduced in one country if you look at like Northern European countries for example versus in other countries where women in particular are you know don't have the agency to make decisions about who they want to be who you want to be or who you want to become is a luxury we take for granted in the western world but it's not a given in other countries right so I mean there's a lot of powered analysis and I'm drifting off like a little bit now from the research itself when looking at this there's a lot of different social science theories about like these gender roles and gender stereotypes and one of the common denominators at least in the social science literature is that it is a construct if you want that like a lot of like how you behave or how I behave and mannerisms that we have are actually just learned features now that stands in opposition to what is generally known as the gender paradox meaning that if if you look on a country level and this is not research that I have done but some of the research that we have been doing is informed by but if you look at the country level why you have certain groups of men versus women like more you know why is there more female physicians in one country or male physicians in another country and what kind of attributes are associated with these professions for example then you have to consider the political system and the social structure where this is coming from but then you also have to look at like the you know freedom you talked about like degrees of freedom like the choice that you have the paradox is not that in Northern European countries where you have generally very high gender equality there is actually very few women that go into STEM research occupations or like traditional like you know more male dominated occupations where you have countries in Albania for example where you have a high degree of women going into these you know predominantly male areas so the researchers have been trying to understand why that is the case wouldn't that like be the case that if you have a country like Sweden where women can choose who they want to be then they would have you know more people in like male dominated areas but I believe it's a U shaped function where we have countries where the women have are maybe exposed to an oppressive system and they are trying to become financially independent or become independent like as a person and they will adapt to a system and therefore also make choices that will allow them to break free where when you have a system where women are free to choose who they want to be women actually it looks like at least fall into like what we consider more gender stereotypical behaviors and that is considered the paradox the point though is if you talk about like larger collective goals then we want to I think I believe it is important to provide freedom of choice and not restrict it in any way like allow people who they want to be equal degrees of economic freedom to self actualize and that's really important and to be able to look at the stream of stimuli that we deliver to children from the moment that they are born into the world and how our culture around them shapes their decisions because many times we talk about men generally being interested in ideas and women generally being interested in aesthetic and then we also and social and women being more interested in social and then we also simultaneously have shaped our civilization around those beliefs so that then it kind of also makes it difficult to in a sense get us out of those ruts like if women do decide to be interested in STEM then sometimes the ruts are so deep that it is hard for us to dig are a way for them to get into that just like with men if they want to get into social work like there is being women in nursing and there is nine women out of every ten people that are around automatically and so again there is these ruts like if I really want to serve people in hospitals that way how do we make it so that it can be easier for people to pursue whatever they want to pursue in terms of having these degrees of freedom and try and limit the social pressures that pressure people into specific categories but rather have it be something that is more like finding out what is really thyself and bringing that forth into the world so that in itself is so with the Twitter data that was being analyzed it was gender biases towards so what we did is that we built a model we first looked at Twitter data and then we quantified the bias in the data and we saw that there is differences in different countries where there is specific concepts that we can measure in language closely associated with female words if you want versus male words very simply put so we had a way to we had a way to quantify gender bias in the Twitter data so we could reliably say that for example in the area of politics or in the area of child care there is more or less bias in one country versus the other so that was the first step and what we then also did is that we basically correlated all this data with statistical indices of countries in order to see if that is actually true if it actually matched the bias in the data actually matches like real world recordings of the world economics forum gender gap index for example and we saw that that is in fact the case so that we in that way we were able to do a check and then in that way we made sure that this is I lost my thread now I'm going to have a yes yes yes we'll take a quick moment to do that I also as you were saying that I also realized that a lot of the things that I've learned in the last couple of years about the field of differences between gender and metrics I would say are very yes of course interesting to analyze the way that we fit into big five personality traits but also just like the sheer complexity of every single person being very unique in terms of their own individualistic expression into the world also I just want to respect the fact that evolution is not something that has stopped and that there's a certain way to frame like this moment and there's constantly a biological and psychosocial evolution that's occurring all the time around us and that these things are never static and so like I don't know something also about just the way that children are funneled into like blue and pink and like like ducks and stuff like that and dolls and stuff like that just it all is that kind of stuff is very strange to me like it's almost it's so binary and non nuance that it's actually kind of scary in a sense like that you can shape your child's life by the purchase of a toy and like that's really scary to me because it's almost as though like it'd be maybe more given the age of the child to be able to enter into the section of toys that they're interested in tools that they're interested in learning from and having them be able to kind of look and try and figure out what interests them the most and then be able to choose themselves I know diversity is beautiful and we don't quite understand that gender diversity is beautiful and not only that it's enriching us and it provides us with a better and not a worse understanding and I think this thinking is particularly important as we are moving forward with the development of AI because what we're doing right now is very often we produce a social reality that is gender biased where we think in pink and blue we have to think about like the values and the perspectives and the knowledge that I mentioned before when we were talking about teams when we are building that AI so that we don't like suddenly export a company like Google and technology into a different part of the world where the local knowledge and the local reality haven't been taken into consideration at all that is very problematic so I think one of the things that we have to really think about and that that we have to think about really deeply is how can we actually integrate the diversity of cultural values when we are building AI systems we are not doing a great job at that if you look at Twitter no no no even in our own research here we looked at men and women like in Twitter data so that's a pretty bad example but we have to of course start somewhere but having an understanding of complex it's also like this hype in AI complex systems it's very hard mega corporations have a big responsibility to I think funnel us towards nuance towards multivariability towards love towards interconnectedness towards harmony towards gathering in person to explore different states of world views rather than simply be dominated by the corrupted business plans of the attention economy feeding echo chamber style feeds and advertisements down our throats we talk about that quite a bit on the program you have a lot of videos already on the program it's very hard to have the time to look at all of that I'm sure you know more we love talking about that space that was let's get into the third one text analysis on how language is weaponized to influence social groups exactly given what you just said about echo chamber I think that's one it's one of the most important things to work on right now is like the social media space as you know you know I shouldn't quote somebody like Mark Zuckerberg but he wants to connect people with facebook and that's what the facebook business is connecting people at all costs so are we feeling more connected thanks to facebook I don't know that's sad we have here a space where people are communicating and for some of us that provides a sense of what used to be a village or the coffee at the corner there's a virtual space where communication happens and there's different influences and different behaviors that are suddenly being displayed there's in cells that have a following stronger than they maybe would otherwise in the real world because there's social punishment that doesn't exist in a lot of these social media spaces so hate speech and cyber bullying and all these negative things that are happening on the social media space have a real world influence and they're not good the goal of this particular research project is how can we understand what's going on in the social media space by looking at the language that is being used in order to understand what the intentions of people are in particular negative intentions and this is important because language is very very powerful and language is intentionally used in order to manipulate and in order to influence others on a large scale it's really a weapon and has become a weapon and therefore it is so important so a lot of money is being poured into this area right now so that we have better tools because there's an ongoing cyber war between countries that is happening right now and it's a very ugly one and the US is not winning it so what we have done in particular it's like a small piece where we're looking again at social science theories in order to inform the development of tools that we can use in order to if we want to find the bad guy but also help people that maybe have troublesome intentions how do we do that there's like a wonderful psychologist called Albert von Durer who used to be at Stanford he has developed really around what we call moral disengagement meaning that we are actually using or engaging in specific cognitive strategies that we are using yeah keep going I shouldn't look at myself yeah keep going so we use particular cognitive strategies in order to justify harmful actions or atrocities if you want this is important because we have to assume and you will like this that in general nobody has you know unless you're a psychopath nobody wants to do harm or nobody thinks they're acting immoral most people think they act morally right even though we might not agree but like I believe that you know my actions are right for me but how do I make them right for me if I actually want to harm you or if I would want to harm somebody else what we do then is that we tell a story to ourselves and that story can be analyzed and that story contains of things such as dehumanizing language or replacement of responsibility because we justify our actions by engaging in these cognitive strategies so that the impact this instinctive feeling that we have that something is wrong is actually lessened this has happened in second world world this is happening in other areas of the world where genocide is justified right now so these are things that we can this is being used by politicians if they want to justify a war in a different country they're all using this kind of language that is full of language that will make the harmful action less impactful when you dehumanize someone you actually dehumanize yourself and that's a really hard loop to understand and it comes from a lot of practice of deep ego loss and ego death and understanding of I actually disagree with you because I don't think that we dehumanize others we dehumanize others because we think we are humans but they are animals they are rats, they are cockroaches they come here and take our things away that we have built so it makes us human but not them I'm still trying to understand the disagreement because I also have that same perspective that you just listed I said that when you dehumanize others you dehumanize yourself meaning that if one gets the idea that you don't actually perceive that to be true you don't think you're doing that even though you are doing that you should dehumanize yourself because we are all humans but the understanding is that I'm a better human in fact I'm so much better I'm human I'm so much more morally righteous that I can make this dehumanizing statement about other people when in reality being the one, being the brothers and sisters that we are trying to collectively prosper on this planet that is a statement that you do dehumanize someone else you dehumanize yourself there needs to be a stronger cohesion between each other where you can parse words that people say on across social platforms and you can predict mass violence and prevent it from happening there's actually a lot of really interesting work happening in what's called data fusion with fusing psychometric data with social data and doing things like predicting violence from occurring and preventing it from occurring with crime prediction these types of things which are quite interesting fields yeah, Palantir is in there a bunch of other companies are now starting to pop up in that field because you can parse peoples for negative sentiment for dehumanizing sentiment and then predict violence and also you can also you could potentially also try and water people's seeds with love because if someone is starting to use dehumanizing language if there is a way to water their seed towards love it can prevent that violence from happening and also it can help them grow more down the line where it's not like we have to put them in prison whether it's they consciously evolve themselves to not cast malevolence on other people that's the preferred solution in the preferred world where is it coming from like if is it desperation or is it anger this analysis that I mentioned like trying to understand intentions sometimes their harm doesn't go towards somebody else it goes also towards oneself in the case of suicide for example so the place where it's coming from where love can help might be easier in one case than the other if somebody feels like they want to harm themselves and ultimately maybe jump from the golden gate bridge love is certainly the right thing hopefully to provide somebody like is full of rage and wants to shoot Muslims in the mall or Mexicans at the border here too I agree that love might be the right tool personally but it has to reach them strategically tactfully it's very hard to when there's psychologists doing a lot of medical profiling on like mass shooters for example and very often there's like specific personality traits associated with narcissism for example that are very prevalent so you have to try to go beyond these barriers because I don't know maybe I don't know enough about love in order to understand how love can not just be given but also perceived by somebody who is so troubled in their maybe narcissistic mind that they would go out and like just randomly shoot people yeah this is going to be a growing field of taking the this extended phenotype of ours now in the digital sphere and analyzing it and making recommendations for our lives that move us in the direction towards love yeah I love you yeah yeah and it's actually right here we talk about that quite a bit when we talk about the ultimate nature of reality or when we talk about things like love when we talk about things like that in the program ultimately it is right here it's right here when you when you sit patiently with someone else and you look at them in the eyes and you you feel the depths of their hearts their spirits are you flirting with me this reality you are a married woman with children but in general we can flirt with each other in the beauty of love by just doing simple things like being present with each other in the depths of each other's psychies and just being quiet in the mind and just being present with each other then all different kinds of very profound things occur where maybe you see some of yourself and me I see some of myself and you that all of a sudden we start merging with our environment together we start taking our drop and merging it with this ocean and then more and more like you start seeing like Sonia as a child her parents their parents their parents you see Alan's child their parents parents parents and you just start seeing different profound things as long as your lens is trained to see the nature of reality when you immerse yourself in these moments I completely agree I think I'm very sensitive to that as well and I'm actually I'm actually quite a shy person or maybe just suffer from a lot of social anxiety I don't know which one but I think it's the most beautiful thing when you really connect with someone when you truly really connect with someone it's so rare it's also very hard to let go of that you want to hold on to it because you feel like there's a unison or you feel like there's a true understanding you don't have to explain yourself anymore that's beautiful but we just need to use that as our food you need to have that energy with us and then ride with it for as long as it lasts because it's one of the most maybe the thing that it's worth living for you have that again with your children as long as you have them of course that's just a very special kind of love and you have that with people that you meet in your life along the way very unexpectedly and that's also a beautiful thing well that answers our last question which is what is the most beautiful oh yeah apart from love I would say I couldn't do without music I think for me personally I think that music will save the world interesting why? it's a creative act not language it's beyond ourselves I think I can lose myself in music and go off into other planets that you want like right there and I think there's just something very deep and moving in music that is so humane if you want that is so connecting that you never feel alone you always feel like something much much much bigger when you listen to good music I should say and that's subjective so that's certainly one of the most beautiful things apart from looking at your children when they wake up yeah this has been so nice thank you thanks for coming on the show thanks for teaching us about social science and AI it was fun awesome thank you for having me there's so much complexity around the subjects that you've been telling us about we appreciate it there's only that much I can say yeah and thanks everyone for tuning in we greatly appreciate it we'd love to hear your thoughts in the comments below let us know what you're thinking and at some point the stream cut but this has been recorded now so do give us your thoughts in the comments below on the episode about all these topics that Sonia was teaching us about talk to your friends, families, coworkers people online about the subjects as well let's have more conversations spurring around the world about the subjects on social science and AI in our future in these spaces check out the links in the bio below to the sift.net profile page for Sonia also hi-c.org also her LinkedIn and Twitter profiles go check those out and support the artists, the entrepreneurs the communities that you believe in support them and help them grow you can find our links below to our show simulation you can find us on PayPal cryptocurrency, Patreon support us and go and build the future everyone manifest your dreams into the world we love you very much thank you for tuning in and we will see you soon peace