ARTS.nano: Fox Harrell, "Narrative, Informatics and Impact"



In terms of my own work Leila just mentioned One of the new initiatives is the founding of the Center for Advanced Virtuality. So this is a center that's for pioneering innovation with technologies of virtuality that include extended reality virtual reality augmented reality mixed reality and more That could include video games social media, but more importantly than that, two new forms unanticipated by any of these And you could say why virtuality my not a center for VR or extended reality and the reason is because One we're interested. Not just in the latest particular technology the latest headset in a way but rather we're interested in exploring experiences that are real but not primarily physical as Mediated by the computer and so what that means is that if you have somebody for instance that is harassed online even to the point of Self-harm tragically, you don't want to say well that was just virtual that wasn't real. And if somebody also has a transformative educational experience or learn something that they can apply in their own life, you also don't want to say that that's just virtual and not real. So in other words, we think about the technologies that you could say paint a layer of imagination on top of the physical world. So there are four different functions to the Center and so I'll just briefly mention these. There's a studio, a lab, salon, and a hub. The studio is focused on making. Now that's new forms of expression innovation. The lab is focused on impact, research, You know, that is the learning, cognition, and really understanding how do we ensure that this work has made a kind of positive empowering social impact. The salon is focused on sharing. You know, that's capacity building events here at MIT and collaborations further afield. And The hub is focused on Involvement. That is, ways we can engage the community and other collaborators within our endeavors. And so now I'll say just a little bit about my own path. You know that sort of got me to this to this point and that some of the kind of works relate to the work of this Center. And so in a way here one of my motivating ideas is this; the AI systems can impact users behaviors, our connections, our senses of self, and even our perspectives of others. And furthermore, since biases can be embedded in technologies, It's imperative to consider their social impacts as in and to do this intrinsically during the act of inventing them. So how can a AI be used for then for social awareness and expression? Well, when I started out doing this kind of work, in the areas of virtuality people actually were interested in something like this. They're interested in creating a holodeck from Star Trek: The Next Generation. That is, a kind of virtual space where you can get a narrative experience or kind of thrill ride type experience, and no matter what you do you get a good story. You know, that is, you can become another person and enter a kind of narrative like Sherlock Holmes and so on. So this is well known and it requires certain kind of technical facility, world modeling Character modeling may be adversarial search so that you can sure no matter what users do then the world ends up Reading I had a plausible outcome. But what I thought instead was could we do something a bit different that? Is could we use AI in an improvisational sense to automatically change stories in the sense of characters perspectives Or what about the cultural worldview or themes or metaphors or emotional tone? And to me that required a kind of engagement with cognitive science What kind of theories of metaphor and how we could generate new ideas and new imaginative constructs and then furthermore? How do we formalize those in a way in that can yield generative computational systems? And so I'll just mention a couple of works so begin Yeah So actually we've done a lot of different a lot of different projects over the years But in terms of the ones so that I'll mention today I'll start with this one, which is called the Living Liberia fabric And so this is a system that was initiated in affiliation with the Truth and Reconciliation Commission of Liberia after 14 years of civil war and the challenge is unlike the more famous Shooting reconciliation commission that came out of South Africa and which you could just point to apartheid and say that okay Well, this was in some sense the spark for the social Ill or you could you could say the shroud of the social ill you know that we want to expose and move beyond In this case, there are 15 different ethno-linguistic groups a series of one-upsmanship and also it was put in motion by occlusion between slaveholders within the United States as well as Repatriated African Americans and became the elite group within Liberia So we needed to think how can we tell this story in a way that respects the multiplicity of the trauma that took place? So used a platform I created which is called griot which is a kind of AI platform to help generate flexible stories that That are different in terms of theme and content each upon each execution. And so I won't go into the details are here I mean it's written up in a number of papers. But basically there's a kind of semantic annotation that takes place and by analogy you find the next best clip to fit the stakeholder group and the theme That you know, that was chosen by the user And So just to give you a little sense Yeah, and so I can't do justice to it with the time here, but just a sense of what this looks like It's a kind of motika fabric. I mean these were produced in Indonesia, but famously popular within West Africa also like a lapa fabric on which people would typically so Images into patterns even of their favorite politician and so on during During rallies and a similar kind of phenomena and in this case we created this kind of work after about a year of fieldwork engaging the aspera Liberians Peace Museum experts and commissioners and so on So it starts off with as ocean sound and all the content on text is generated kind of poetic text And you choose a stakeholder group at the beginning based on that stakeholder group. It'll put into motion So we chose here After You Now all that that footage was collected in the field in a kiosk using a system with agent to help people can work Facilitate people weren't necessarily printer computationally literate but of course management and many other animes and telling the stories that first part of my life was angry with the perpetrators, which meant these exchanges when I started working with them, I realized that these young people as as much victim And so as she spoke about the child soldiers we have another set of potential Clips here and So I'll stop here here with just you know That work we can engage it more if you like and just move on to another system Which is using another platform of our design This is called grayscale that addresses sexism in the workplace and it addresses a particular Model using a different kind of immersive system. It's an epistolary narrative. Now. These are narratives like they'll take place over journal entries or some series of kind of text in this case emails and It makes you feel like you're a HR manager in the face of In a kind of melancholy great corporate environment We have a model on the backend That's using our system called chyme area that models membership in different groups in this case different types of sexism and then you have to choose how you respond to these kind of incidents such as Here, I mean, this is just one but here Rob is written to the company list our company has a dress code for a reason yoga pants are a Unprofessional and be distracting. All right. So what do you do in the face of a series of messages? You know that are I mean, this is just one. We're actually I'll say a little bit about the underlying model so do you then say something like it is not your policy or not your responsibility to police others clothing or Do you say something about the fact that these are being called distracting is it what in fact is the issue here? and so you have a series of different kind of responses that are aligned with different types of Sexism. Yeah So this is implementing a social science model of sexism That includes both the kind of overt hostile sexism as well as complimentary gender differentiation That is like well, you must be good at the kind of people skills. You know, this this kind of these kind of issues or protective paternalism and so on and the system is tracking the trajectory of your sexist or non sexist responses and generating a host of different kind of eventualities Then one day I'll conclude on is a system that some might have heard about because this was On display at the MIT Museum has been shown in tel-aviv and France and many other places And this is a system. The director actually was a visiting artist through Castor wrote a grant to bring him here he was also in my lab the imagination computation expression lab and Cast was very generous in terms of supporting this set collaboration So this is a work that puts you into conversation in a room That's about as big as the front of this auditorium like a hotel bar room but you're in a VR space you have combatants on either side of the room and those combatants end up responding to your body language This is the part that I've worked on your nervousness bias and so on and changing the story and stage Staging lighting and even at the end you look like the combatant that you were at least comfortable with as you went through the experience And so you have a sense of what the space looks like has three different conflicts has gang gangs in El Salvador Conflict in East Congo and Gaza and you always get the one that you are least familiar with at the very end based on kind of pre-registration and So I'll wrap up just with a brief trailer just to give you a sense of work and then we'll have our next two panelists – really? The enemy was born out of my frustration as a photojournalist and more correspondent For almost 20 years. I have photograph conflicts and witnessed the consequences of huge geopolitical shifts When I became a father I simply knew I could not keep working on the front lines Yet I was not done trying to understand Wars Oh never workers, you know about a New York My friends in Israel when they know I'm heading for Gaza come help themselves but to wish me luck and to stay safe They believe a lot of people in Gaza I irrational Also when I spent weeks working in Gaza, and I'm about to return to Israel My Palestinian friends are telling me the exact same thing. Be careful there The project is rooted in my experience as a war photographer going from one side of the frontline to the other and Finding that the fighters dreams hopes and nightmares are far more similar and they are different So there is a bigger story than the war itself and this is the one I want to explore and share For the enemy I am using the latest technologies in virtual and augmented reality So you can engage directly with the combatants and meet them hear them and feel them the way I did In many part of our walls you create an enemy as a kid without having met your enemy Because a society around you has created an enemy in the other So the question is could I be you if I was on the other side So it just as I started Then we could conclude just thinking how AI systems impact users behaviors connection senses of self perspectives of others You know, we've had a few examples of this now and then since biases can be embedded in technologies It's imperative to consider and design for their social impacts Thank you very much You

Be First to Comment

Leave a Reply

Your email address will not be published. Required fields are marked *