Barnard Computer Science Seminar - Alexandra Kitson - Simon Fraser University - December 12, 2024
[Rebecca Wright] Good morning everyone. I'm Rebecca Wright, chair of computer science at Barnard. We're here for our computer science seminar today and really excited to have Dr Alexandra Kitson with us today uh from Simon Fraser University where she is a multidisciplinary researcher at the School of interactive Arts and Technology Her work focuses on designing developing and evaluating emerging technologies such as XR, AI, and wearable systems to promote health and well-being and has published in top venues as well as received awards from things like ACM CHI, IEEE VR, and NeurIPS, and she will be talking to us today about her work addressing real world applications through community engaged research. Alexandra the floor is yours. [Alexandra Kitson] Great, thank you. Okay so hi everyone today in my talk I'll I'll talk a bit about me and my background because I think it's important to understand where I come from geographically um and also my academic background and how that has shaped my research and teaching I'll talk a bit about the research questions that I'm interested in uh the theory that I use in my work uh I'll do two projects just to highlight the the kind of breadth of research that I do I'll highlight some student projects at the end and some of the teaching I do and some future directions where I see I could be going towards at barard so I'm from the, as Americans call it, Pacific Northwest which is just like the west coast of Canada uh this is from Vancouver Island in British Columbia and that's where I grew up so just surrounded by nature and trees and small communities so my town had 2,000 people uh really small and um I went to school in Vancouver my background is in Psychology and computer science and philosophy so I was trying to bring in different disciplines already really early in my career and then I went to the school of interactive Arts and Technology for my PhD where I focused more on human computer interaction HCI and design and trying to bridge all my different interests together great minimize sure I can do that that okay Peaks as well okay great all right all these Tech things um my motivation so I used to volunteer at the Vancouver crisis center and suicide hotline for seven years where I took calls and also had an online chat talked to folks from uh the Lower Mainland of Vancouver uh also youth from Canada wide so we were one of the first chat services in the country to talk to Youth and um coming from the Canadian perspective there's a lot of uh lack of Access to Health to mental health care so although we do have access to physical Health Care Mental Health Care is not part of that so it's a huge gap in our system and there's a lot of systemic inequalities where folks um don't have the means to or it's so stigmatized in their community that they don't feel they're able to reach out and get help and it's a huge burden also on our economic system um especially in Vancouver where it's a kind of milder climate we have a lot of people who come there looking for services and end up on the streets so I was really seeing this um uh day-to-day in my work at the crisis center and I wanted to use my skills in Computing to do something about it so the primary research question that drives my work is how can we design develop and evaluate interactive Technologies for social good and wellbeing so how can we connect people to each other to strengthen our communities and use technology to give us more access and resources to different Mental Health Technologies Yu so what does it even mean to be you know what is mental health or well-being in a technal technological world when we're surrounded by this Tech um even today we all have our laptops and phones and things what does digital well-being even mean and so in my PhD I I set out to to try and understand what it is we're talking about and I looked at a lot of different areas through economics through engineering through psychology through HCI and a lot of these were all saying similar things just like in a different way coming at it from like a different lens and so in this paper uh this review paper in Frontiers I tried to bring this all together into some kind of conceptual framework that designers and developers could use as a means to ground their um development in and part of that was um how I think of it is bringing in the theory from Psychology from HCI uh even clinical Theory and and bringing that together with practice-based experts so always grounding it and lived experience of people and working with them uh as well as the the communities where this Tech is deployed eventually um so it's really taking this like triangulating and and taking these three and bringing those together where we can come up with solutions that won't work whenever I set out to do a project I think of what is the logic model what is the theory of change especially for something like mental health I think it's really important to understand the underlying mechanisms that lead to the behavior change or that leads to a clinical outcome before going in with technology otherwise and I see this a lot in HCI especially we create some tech which we think might help people it might give them positive emotions or something but it's kind of missing this logic model it's missing those pieces those dominoes that line up where if you bring in Tech into one piece you know it it will work um so here's an example of going input process output outcome and eventual impact where you could develop like you want to develop an app to improve people's Fitness and so you start with people those are the input in the mobile app and they're not really into fitness bringing in the app you have these activities and engagement features like a calendar uh which could work with their schedule and and slot in different times to exercise that works for them the output is doing the exercise a scheduled calendar comes up it's like okay time for that spin class and then the outcome is okay now people are being active more but then impact is like over time if they keep doing this then eventually that can improve Fitness so this kind of logic model this top part can really work for a lot of different uh design spaces especially around health and wellbeing uh in my paper we were really interested in like feedback mechanisms as well so here at the top that's the theory going in to guide the interaction design and and strategy elements but you have here this is the input just like back here these are the this is the system the interactivity and processes that are Guided by Theory and then the output is the user Us's immediate state so you could have some positive emotions in the moment while using a technology for example a breathing app which tells you helps you breathe more slowly helps you feel more calm in the moment and then over time you can see that it can lead to uh different impacts like feeling more present at a meeting feeling more connected to people so that's that's one of the papers I've done uh and then how I use that theory in my different projects I have quite a few projects if you've looked at my website before um it's quite uh the breadth of projects but also depth in specifically virtual reality in wearable tech so I started doing virtual reality in um 2012 you can see this is the the Oculus now meta before they were Acquired and our lab um did the kickstarter for them and so this was super early days when we really needed to figure out the foundational research and um doing a lot of the the hardware of it like we didn't even have controllers so how do you move around this virtual space so we were figuring out different Locomotion systems we were also trying to figure out because a lot of people were getting sick sick quite quickly trying to figure out different ways that we could mitigate that um through the hardware and software and the design um and then in the VR set and setting that was also about like exhibitions or we were noticing if we just put people in a headset and then take them out they were getting very disoriented and that was because they were so disconnected from the reality that we share right now and so we tried to uh use like museums as an example as a way to slowly bring people into the virtual reality and transition them back out again Sonic cranel was another one which wasn't uh like immersive visual VR but it was a virtual soundscape so virtual reality isn't just visuals it's all of our different senses and audio is is one of them we explored where people could change their breathing which would change the soundscape around them while they're sitting in a nice hammock uh after some of this uh kind of foundational work on exploring V I went into some applied uh applications uh this first one is on how we can connect people in virtual spaces and so um you have you can have two people in the VR headset and they don't have to be in the same room at all but in the virtual space they're connected through the track sensors uh and the cameras and the headset and they're these particle bodies which as they move closer or further the way they change colors and shapes and react in different ways so that was quite a playful thing and we demoed this at like the Vancouver Film Festival um and the Richmond uh World Expo so at we also do a lot of like public exhibitions in my work Odyssey was another one that was uh around connecting to the planet and we wanted to have people think about climate change and how really we're all trying to save our world and planet and and so Odyssey was designed around that intention where uh you would feel connected to the Earth and astronauts experience this when they go into space it's called the overview effect and um we tried to capture some of that by interviewing astronauts and then we worked with a crew of cosmonauts who were training to go to Mars in an isolation study where they're like trapped for six months in a bunker somewhere and we gave them this VR experience as a way to um connect back back to the planet and remember the intention of their mission uh we also did some work around Altered States in VR and trying to look at the parallels between different Altered States Of Consciousness and different realities in VR I thought it was quite interesting how in virtual reality uh you're you're kind of aware that it's not real but it feels quite real um almost like a dream and so I'll talk a bit about that one um but for now um a little bit about these other ones on the bottom that's my more recent work so in my postto I looked at the ethics of bow wearable tech by working with middle schoolers uh during covid we ran an online Workshop where they assembled this kit and then um through video conferencing we scaffolded the learning process of uh how to how you can design this but then also had created these um ethics design cards which got them to reflect on uh the design decisions that go into these um products and how that could impact their sense of self so one example was this uh one girl and and she was part of this Workshop but she had a lot going on she was on her way to basketball practice and so she's uh on zoom in our workshop with her kit but in the backseat of the car on way to a basketball game and she's like uh pushing these different changes into her uh little wearable kit here and there's this little LED display here which changes in different ways and initially she had it so there's this bar that goes up and down and you're supposed to match it with your breathing she was doing that and I said well how does that make you feel do you feel ready for your basketball game she said no like I can't match it I'm feeling so anxious so then we worked through how to change that so it could work more for her and so she changed it into a purple disco ball that was just glistening and she was just breathing and watching that and that made her a lot more calm right before her game so that was cool we've also done some work in uh high schools and I'll talk more about this project too later in my talk uh where we go in and uh talk about emotion regulation and then how virtual reality could be a way to help with skills development and then these last two are on kind of going back to my initial ones like how how can we develop the more foundational aspects of this technology and I'm really excited about social virtual reality so now we're seeing devices that can do different like personas you might have seen meta and apple have these mixed reality personas that can be in your space like right right in front of you and you can talk to this person like they're there um and then also different devices so having cross device so if you don't have a headset or you don't want to wear one you could still interact with people who are like through your phone or your laptop so in this project we were kind of exploring how that could look like in the future and uh also for language learning we looked at Social virtual reality is a space where you can practice language learning with people uh but then um we're also curious in how llm agents who have gotten a lot better in recent years could be another tool for that okay so I'll talk about these two projects so the first one is lucid Lo which is a neuro feedback augmented immersive experience uh Lucid Loop is inspired by lucid dreaming and you might have lucid dreamed before so that's when you're dreaming and you realize that you're dreaming and then you could just kind of watch your dream happen you wake up or sometimes you can even control your dreams so you can potentially do whatever you want in your dreams and I thought this was a very interesting experience that I wanted to do more of it's like the ultimate virtual reality but also I thought well maybe some of that could inspire a lucid or a virtual reality experience that could give people kind of insight in how to do this so uh I sought to to look at what are the important design features that should go into this experience um there were two main skills to have more regular and sustained lucid dreams that we already knew of so again this is the theory going in the theory of change and that's practicing focused attention and uh pneumonic induction skills and that's just like a memory aid a lot of people will throughout the day just kind of pause and maybe look at their hands look at their watch look back again these are kind of like reality checks to see if you're Dreaming or not um some people pinch themselves Am I Dreaming uh so these are kinds of pneumonic devices and then um in what ways are lucid dreaming and VR kind of alike and to similar can we learn anything about it for design so um looking at the lived experience of people so there is theory and now it's a lived experience uh I interviewed expert lucid dreamers from around the world and they had 20 years of lucid dreaming like quite proficient and we looked at a phenomenological analysis this a kind of qualitative analysis to really understand what that experience is like and the things that lead to them becoming more Lucid or not and we translated those into uh different design considerations for Lucid Loop and that was um like it should have like a painterly aesthetic dreams for them were often quite realistic but then sometimes at the edges or or they would get lost in the dream it would feel kind of surreal also quite Vivid in terms of like the the quality of a dream some things feel really sharp While others uh are dull more in the background depending on what you're focused on um we brought in particles forest and some scarf dancers because again these were kind of natural settings plus a surreal element to it which we thought could help people practice that kind of oh this is a dream this isn't actually real we also added in audio like Whispering as a pneumonic device uh to tell them you know next time you're dreaming you'll have a lucid dream uh and it was on a loop so we shot this video on a loop so you could keep playing it for however long you wanted and it would be different uh we decided to go no Avatar in this one some people dream and they they can see themselves or even from a third person um but a lot most people in our interviews uh didn't have a body and then different levels of abstraction so we we had different layers of realism and surrealism in this experience this is some of the design process we did so we took a um uh Steve dea's deep dream uh painterly system and we ran our 360 video through it and tried to look for you know a painterly aesthetic with Vivid colors that seemed like it go could go between real and surreal and some of these are quite you can't even make out what's happening where others are are more realistic and so we had to choose from over a hundred different ones like this uh the input we decided on was a consumer EEG headset The Muse 2 and it just looks at U the electrical activity on your on your skull which is associated with different brain States and I just want to say it's very Loosely tied to different emotional states like attention it's not perfect and that was okay for our experience because we were going for a more artistic angle uh it wasn't meant to be any kind of like medical grade device and we also went with a head mounted display so that people could be really immersed in the visuals rather than like a projector system um and kind of block out and just focus on the experience itself we had our deep dream visuals and spatialized audio as well so the loop kind of looked like this where people were seeing and hearing different things in the environments and then it was it was looping up to this uh Muse monitor app which was taking the the EEG data and then we using osc put that into uh the Oculus a Unity program where then we had our own like algorithms to filter it and um turn it into a signal that we could map to the visuals and the sound uh in real time the only thing that wasn't real time was the actual like visuals changing we kind of had to fake that so we created seven different different layers um because we made this in 2017 and some of the generative AI stuff was not quite where it is now but you can see there the layers going from one to seven these are just some samples where if you're more focused your attention the visuals are clearer the sound is clearer just like lucid dreamers and then it gets more abstract more you lose focus but then there's also some things in the environment for people to kind of pick up on and regain their focus we also added some computer graphics on over top of this so there were some like particle systems and uh different little animations that could pop up in and out as a surprise so here is uh a little video kind of highlighting what this experience is like with um some audio in the middle we'll see if the audio e [Music] [Applause] [Music] I turned it off I'm Dreaming next time I'm Dreaming remember remember I'm remember [Music] right so then after creating this based off of the expert interviews and some iterative design we looked to um some uh user experience research and to see if the thing we made actually is doing the thing we wanted it to do so we use a thing called cued recall debrief where it's a little bit different from having people talk aloud as they're using uh product or a system we get them to um put on some they can put on some sensors to capture physiological data or not in our case it was the EEG which then also ma to the visuals uh but we just had them run through the experience um without the need to think about like how they're going to tell us about it or comment as they're doing it because I think that really takes you out of the experience for some things and then instead play them back a recording of what they just experience and then they can narrate that as they go through it so there's um you know it's always tricky to kind of balance that where you don't want to um have people look retrospectively at something because sometimes you forget what the thing was after you experienced it but you also don't want to take them out of the experience so that's kind of a an interesting different way to do that and we also did some semi-structured interviews and um questionnaires to get at their experience so I really try and take like a a multi-pronged approach to understanding something and measuring the outcomes so some of the themes were it was kind of like a psychedelic experience for some people um where you go into some people felt a lot of fear and uh they didn't know what was happening their minds were going left right and Center and they were seeing that reflected in the visuals and sound and some people said that uh rather than trying to freak out about it they remembered like oh actually if I just surrendered to it and relax everything's okay there's also a sense of Discovery so you know you start to focus or concentrate things zoom in on patterns or zoom out things get louder or quieter and for this participant it was cool because they tried to do a certain thing and it did did it as expected so they're getting that feedback so for us it's like okay our system is kind of working as intended uh even though when I go back and look at the data and see like their like their brain States and and how that's mapped like I'm not sure it's exactly mapped to what they're thinking but for them they were reading into it um so also just yeah being there for them uh it felt like they had some agency over the scene I thought um one participants said that there is a path and they wanted to go down the path and they were making up uh different stories about the characters in the scene you know just us researchers but but for them they were um putting their own interpretations onto a very kind of abstract environment so I thought there's something there that's cool with the second research question you know how are Lucid Loop and lucid dreams how is VR and lucid dream related the sense of fluidity was definitely something moving through the different types of visuals and sounds uh everything is changing constantly like in some dreams also the emotionality of it so a lot of people you know had fear it was on of them but also like peacefulness calmness connection to the forest and also to the people a lot of people were questioning reality so some people uh in the VR headset where we're thinking like is this really doing what I I think it's doing um is this what my brain is doing right now um so they were kind of questioning reality in a way that in lucid dreams you also question if the dream is real and it was also very quite reflective so arant actually wrote to me after it said that that experience um made them think about different things in their life and also uh helped them become better lucid dreamers as well so I haven't done a follow-up study to see if it actually induces more or persistent lucid dreams but that would be a future study I'd be interested in so this resulted in four papers at top hcii venues like ACM Kai where we demoed it there also gone to MIT media lab where we did a dream engineering Workshop which um spawned a lot of different collaborations and uh I'll show you one of the the student projects at the end that that also came about um and recently got a grant with a couple dream a Engineers who work more in like the psychology where um we can develop a system that is using more generative AI rather than the Deep dream that Luc was it was also exhibited in Scotland and in France um also locally in Vancouver uh some different collaborations with researchers training students and all of the the code that we created so the the filtering of the EG data and mapping that to the visuals and audio is all open source landed on GitHub and also was funded by my uh social science and Humanities grant for my doctoral studies right second project was VR for emotion regulation youth so this is a little bit different maybe a little less artistic but now more applied to uh a certain mental health problem this was my postc it was very busy I got to say um we did a lot around the theory trying to bring in different experts also um end users with youth we had three different workshops that we ran over a year an Advisory Group that we frequently checked in with um and now we're kind of at development and field study stage so still ongoing work and as I said earlier youth are one of the particular populations who are most at risk uh for having Mental Health challenges and in Canada suicide is the third it's the third highest uh in the industrialized World wait times for counseling and to get help are really high and this is a really pivotal key developmental stage for them where learning emotion regulation skills is so crucial if you learn it then then those skills will really serve you later in life so that's why we chose them um we're working again bringing in theory with James gross's uh process model of emotion regulation where uh he says that we identify an emotion we select a strategy to change the emotion we implement it and then we check in ourselves to see if anything changed if not we might go through several Cycles until we get our desired result and there are different ways of regulating our emotions and you can categorize them like this but they're not strict categories one of the ones I was most interested in is cognitive change one because it's one of the most studied but also it's one of the most challenging especially for youth even us we're still developing these skills and cognitive change uh also has different types uh and it's really if a situation has already happened and you're feeling the negative emotion uh you can use different uh cognitive strategies to kind of reframe a situation so the situation is actually the same but the way that you relate to it or think about it is different and that can make us feel differently for example it's feeling this way really that bad or I feel bad but it shows that I care so these are strategies that teens are still learning and we looked at different psychological therapy manuals so like over 40 therapy manuals we analyze to try and understand this cognitive change process and the challenges uh within but we realized also clinical practice doesn't really isn't really captured in these books so then we went to clinicians and interviewed them to uh more deeply understand the challenges so then once we knew of the challenges I you know we're seeing some opportunities for virtual reality to come in and virtual reality I think has three things that are really unique about it it can immerse you your senses you can feel like you're really there and you can also be embodied so you can have an avatar and you can look around pick up things and it feels like you're with someone or really there in an environment um and in this way it can really evoke a visceral experience emotional experience in a way that feels quite realistic and you can simulate uh different experiences and try things out over and over again without the fear of actually failing in real life it's good for teens so we created these prototypes we'll see how this video goes um and these are are different kinds of cognitive change strategies that are then running I want to R it with you a meeting in this room we will try to uncover the emotions of the green player emotion starts with Sensations in our bodies purple player please press on the emojis that you have placed to pick a more nuanced emotion word can you resize the emotion balls to fit how much your friend has felt these emotions [Music] [Music] [Music] [Music] [Music] great so we created these four different prototypes it was a team of four undergrad students and uh one PhD student helping with the project management and so they came from different backgrounds there was a psychologist there was an engineer uh there was a computer scientist and U there was an art and design student and so I was able to lead them and we they all came up with these four different Sprints which they did over the summer so each was only two weeks and they were able to um uh take the the theory that I had T them and Implement them into these different very small prototypes which we then could show students in our high school workshops as like potential things you could do with VR so yeah they're quite like low Fidelity prototypes but they do the job of showing like what's possible we also use like tilt brush to uh incorporate it into our project and uh yeah I think that they learned a lot so then we took those prototypes and we went into high schools where our goal was to identify opportunities for VR to support their skills development uh the kind of challenges they faced from their perspective so not the clinical perspective and how they envisioned VR could help them we had three kind of learning objectives where we wanted to sensitize them and upscale them on emotion regulation as well as virtual reality features we also wanted them to have some body based reflection because VR is you use your body a lot you're moving around it's 3D so we want to see like what could they use that for for emotion regulation and emotion is so embodied as well so they seem to go really well together and then look at design fiction activities or scenarios as a means to think about the challenges so these were our three workshops and three different classes they from different diverse backgrounds um uh a lot of the low sces high schools were from like immigrant families and we just tailored the activities around their kind of expertise so for this one it was more about like doing brain storming in some ux different design activities whereas this one was drama students so we got them to do their drama activities and we worked a lot with the instructor too to make sure keep the kids safe and us safe and um use some of her drama activities as part of our design collaboration so some of the highlights from the different workshops so you can see these ones we got them to do like rock paper scissors with their bodies that was one to get them more into their bodies we had lots of discussions whereas this one we had similar activities but the differences here where uh doing drama activities having different kinds of demos and then acting out how they imagined VR could be as a skills development tool for emotion regulation in the future these were some of the challenges that they came up with so one was inflexible imagination sometimes it's hard to understand another person's point of view so VR could maybe be that they also thought that um having a social experience would be really great because a lot of these emotionally charged situations are around their friend groups they also like the idea of externalizing the inner self so a lot of time there's a lot of stuff happening in our bodies and our emotions and it's hard to deal with especially for teens emotions are Amplified so if you can get that out there sometimes that distance can really help you see the bigger picture see cognitive reappraisal yes the cognitive change process so we took these ideas from Youth and then we also did a scoping review of existing HCI interventions to see okay well this is what teens want and is important to them these are the reappraisal challenges and these are the VR features they're most interested in where can we see some gaps where HCI is not covering that that could be potentially viable and we looked at um cognitive reappraisal challenge themes sub themes and we saw that um yeah being able to visualize what's happening in our bodies having social experiences and also personifying our emotions was uh interesting Avenue in the paper I did at Kai this year I came up with some just design fiction scenarios of what I imagin this could look like and so those you know people can take them and go with them as they will but I also want to explore that more in future work um so in terms of impact we did 78 teams from diverse backgrounds five teachers Incorporated their uh our Concepts into their curriculum also inspired multi-disciplinary research projects with I'm now working with BC Children's Hospital and trained a bunch of students and had them great Publications so really trying to have impact inside and outside Academia I think I have a little bit more time to show two student projects so I'll just talk about maybe this one uh future Earth event this was for my class which is environments it's a undergrad class um interactive Arts the previous anime and there are teams of there's a them climate awareness so in this one in the game in virtual reality you have Wasteland Rob popular how you do this for um they have seven weeks where they're learning and doing assignments and then seven weeks to do this and then there's a project showcase at the end where we invite industry to look at [Music] it was just pause it here because it shows so in this one it it wasn't really about um I think the theme was do something unique it was very Broad and they played with the um a short and tall like scale in VR so the player had to like change their scale in order to complete the puzzles that was kind of fun uh and then grad student projects so this one is actually today at I couldn't be there and it's around using virtual reality speaking into your hand and there an ear there and you tell it your dream that you had and then that translates it to texts it into an llm which then creates these 3D meshes in virtual reality that you can um play around with and scale and reflect on what these dream objects mean to you Elma T is a gp4 based embodied conversation yes it's an embodied conversation movie d um who looks like this and you can go into social virtual reality it's VR chat it's a really popular social VR platform and talk to this uh agent and practice your at English skills so we ran a study with that Elma T is a so yeah the kinds of future directions I'm looking to go in are are still kind of in this like Health applied space but also in like fundamental um virtual reality and computer graphics uh Direction where like how do we actually represent people in these digital spaces how can they interact um what are they doing there how how do we represent them and then how can we then co-create personalized mental health interventions I'll leave it there thank you all right uh so we can take some questions uh either in the room or on the zoom if you're on Zoom you can uh raise your hand and we'll call on you uh and figure out if we can hear you when we do that um also you can type your question into the chat and uh we can read it out for you questions question so I think this is actually super interesting if you imagine and I have people kind of um looking a bit of the public house and your epidemic and then a lot of people that are actually jails and kind of having this experience of VR replacing that all of that and having the outside experience so imag this type of Technology how would you imagine that being because they really cannot have interaction side yeah yeah absolutely there was a project which unfortunately didn't go forward because of Co but it was um uh Ju detention center of young girls and we thought about bringing in virtual reality as a way to practice um being in triggering spaces because a lot of the time when they go back to their you know they they learn they get treatment they then they go back to their environment and then Things Fall Apart so how could you simulate that as much as possible in VR and then we also thought about it as a as a mediator to talk to certain people like family members or friends from a distance yeah I think there's huge potential there more I'm really interested because I on like L side project on learning like helping people using V art help them learn yeah so in that case um like with the last project the student project I showed with the llm yeah I'm I'm super curious about that and how these agent these embodied conversational agents could be like teachers or people we could practice certain skills with be it language or interviewing or onboarding people or like an assistant where you're situated in a certain space um and they can be kind of like your guide I think I did see somebody last year did like if you're at a party and you're kind of nervous you don't know how to interact with people you could have this embodied conversational agent as your kind of wing person I guess so yeah it's possible I think but um what I was Finding especially with the llms is after 10 minutes or so things just they're done I actually curious to see they about 10 minutes and then things go weird but yeah I wonder like if you can bring in different models so maybe they could switch at a certain point and they don't get too tired or whatever's happen yeah um I guess I have a question about that as well so in addition to the sort of going weird or hallucinations even at their best it feels like current llm chat Bots are fairly generic like they're l is fairly generic their style is fairly generic um I just wonder if there's a worry if you use them as teaching tools where we're sort of going to spiral our all of humanity into this genericness where it learns from us and we learn from it and um well you can even see in like people's writing too there are certain phrases that that are now being more prevalent which we don't normally speak like so I think yeah it's very important to think of like yeah what are we putting into these models and maybe maybe we don't need like an llm necessarily but like smaller language models where we know what we're inputting into these models and could be more helpful yeah yeah yeah I thought it's not I'm trying to formulate a question but maybe I'll just say it aloud in question but I thought the context you were describing of like social interaction seems um really interesting um from sort of like the mental health standpoint and I was curious in your conversations with clinicians or um psychologists like where they kind of land on the spectrum of like virtual interactions as opposed to interacting with others people within like other real people I guess with a virtual space because um on the one hand I think you brought up this idea of like repeatability like you can have this sort of safe space to practice things and maybe virtual avatars might be better for that but then there's also sort of like the ethics of you know when you're interacting with something that is um sort of like virtual what does that mean in terms of people learning interactions when there might be limitations around like how fluent or expressive they can be and how much they relate to maybe conversations you might have with your so just like where how did they think about that conversations with experts yeah it's a great question and they're definitely cautious about any technology like even using a phone like they're very very um traditional in that sense P to face versus like or yeah like yeah even using yeah like remote counseling for example I think they would prefer face Toof face in person but sometimes that's just not possible and sometimes people have like um phobias or things around being in person or talking to people one onone so they said they see the potential of this technology as a kind of Step stone or one tool that they could use and they're already Big Tool Box where people could like a first introduction kind of like exposure therapy it's like the first little exposure to practicing or with teens um you know they don't want to take home a worksheet for example to you know write down your thoughts and what you felt and it's like they're not going to do that but maybe if it was like a VR game where it's like built into to something fun they might be more inclined to do that and so that kind of retention might appeal I guess to clinicians yeah follow question actually I was thinking the news I saw something I think it was this week around like a social robot for kids that was getting discontinued did you see this no I'll you over this a longer question so save it for that okay sounds good all right I think we'll and there uh Alex thank you again thank you all right thank you guys on Zoom see you later bye
2025-01-05 05:49
