The final presentation for the conference, final presentation for the whole year, so to speak. So it's fitting that it's by our friend and co-organizer, Jan. Yes, that's up to Jan. Thank you. I don't know if I... Okay, should be okay. How do I use that?Doesn't... Ça va, mais attends, je fais comment? Ah, ok. So, good evening everyone. I'm Jeanne Lechevalier and I'm going to talk about the topic here, the face tasted by artificial intelligence or the face put to the test by artificial intelligence. So I want to first say that I started my research with the fact that a visual landscape, our visual landscape today is saturated by faces in advertising, cinema, in social media, in dating apps, if you tried. And it's... We are always facing faces somehow. And technologies are more and more implied in our relationship to these images of face. So I want to first...like see with you different dimensions of the face. We have a guest. I don't know. And also just a parenthesis to tell you that I'm not an English speaker, so I don't speak English very well, so maybe sometimes it will be a bit strange. Don't judge. So I want to talk about the different dimensions of the face to see where could be the problems. First, there is this appearance, the physical appearance. So that's the first way we apprehend the faces in our lives. Like we see each other, we judge each other. And that's more and more the case when we, with all these technologies, you know, advertising, cinema, we less see the faces directly, but more through some medium, you know. But the face is also the manifestation of the presence. So it's a kind of transition between the exteriority and your interiority. It's not just material, it's also symbolic. It shows a part of who you are, but at the same time it hides another part. AndWhat about AI in all of this? AI can kind of manipulate the presence. That's the main problem that we face with AI today. It means that, for example, with deepfakes, I don't know if you know this technology, but it allows you to use the physical appearance of someone. And to make it speak or act in a way that he never did, without his consent. So it manipulates you about the information it gives you about this person. This person never did this thing, but you think that she did or he did this thing because you saw video with this technology of AI that shows this person doing something. Well, I'm going to be more precise after. But, so, yeah, so the AI is not the problem in itself, but it's our use of AI that could be the problem. And so the core question is, can the ethical face survive its aesthetical reduction to a manipulable image? So can we still conserve responsibility, I don't know, kind of respect, consideration of the other if we cannot trust the other anymore?because of the use of AI, because we can manipulate all the images and manipulate your identity. So first, I would like to, in the first part, explore with you the face as a technical surface. So technical surface, it means that we're going to take into account this first dimension I told you about in the introduction, which is the physical appearance. So we can see in the art history, because I have a double format training in art history and in philosophy. So we can see in art history that the portrait is a way to represent, not to reproduce. And there is a big difference. For example, you have here different images of the same protagonist or person, which is the Christ. And the Christ, so Jesus Christ has been represented as, for example, a fish, as a phoenix, or you can see here as the Pantocrator, who is the kind of omnipotent God. But you also have the Echeomo. Echeomo is the suffering Christ, the ones who is hurt and vulnerable. So you have different representations of the same person. And there is no manipulation in it.All the representations are legitimate. That means that the portraits and the representation of a face can be symbolic and not always try to copy the appearance of someone who existed. And you have another example here, which is the scream from Edouard Munch. And the scream is an interesting portrait because it doesn't portray an individual, but it portrays a universal emotion, which is anguish. So the face also has a partially universal meaning. We are all different, but we can feel or find the same feelings in the look, in the face, in what we feel in our interactions with someone.But I'm going to be quite quick on this slide. We have maybe a bit, I don't know if it's more and more, but with the technologies that we're using today, we have an instrumental use of the face. It is less used as a way to represent symbolically something that belongs to every being, but it's more used to completely detach from the individual. So I'm not going to talk about the restoration part, but I'm going to focus on the decorative turn, when the face becomes a sign detached from the individual. We can see it with Marilyn Monroe, for example. So I took the example of the diptych from Andy Warhol. In the 60s and you can see where he took the original photo which is for the promotion of the movie Niagara. I don't know if you ever watched it. But well, in this diptych, we can make the link with what Walter Benjamin said in his research, that there is a loss of aura.That means that the image is completely disconnected from the person behind. We're not seeing the actress Marilyn Monroe, we're seeing a product. She became a kind of, I don't know how to say it, merchandise? Well, something that we can consume. And not a human anymore. And this instrumental use, we see it in different examples that are, I would say, exemplar, like today, I mean, that are showing you how it's very strong now today, this instrumental use. So through advertising, for example, we don't see very well the images, I'm sorry, but in the first photography we can see Nicole Kidman with a bottle of Schweppes and her face is used to highlight the bottle of Schweppes. And it's like shows, it's kind of make a link between her beauty and the taste of the Schweppes and the fact that you should drink it because there is someone like Nicole Kidman who's drinking it. And also a way to instrumentalize the femininity, the beauty, things like this, to sell a product. So it's a way to seduce you.And you have the physiognomy, which is a very interesting example because the physiognomy, it's a science that comes from the antiquity. Okay, we can find some references in the Aristotle essays. And this is the science that says that we can read what's inside of you based on your personal traits, you know, on your face. Okay? So I put an image from the Charles Lebrun books that are from the 18th century. This is the moment where the physiognomy started to be very famous with the work of Lavater. And there is an association between the face of animals and the human face. Like for example, if you look like a lion, it means that you have a strong character or something like this. If you look like an eagle, for example, it means that you're probably a bad person or an aggressive person. I don't know. And today, an example that is very recent, so deepfakes, as I told you, and simulacra. Here I have two examples, but I'm afraid to take too much time. Well, the first example is about Taylor Swift. I don't know if you heard about it, but in...like two years ago, approximately in 2024, we found like deep fakes of her, so images of her, but without any clothes, like completely undressed, naked, or in a sexual context, okay? And it had an effect of, how to say that, destroyed her reputation, I mean not destroyed, but still had an effect on her reputation for a moment, and she had to clarify it and things like this. So this deepfake, the way to use the face with the technologies had concrete consequences on her life and on the life of many persons who are victims of that. And there is another example that is very interesting, I would say. It's the Jessica Foster example. So I don't know if you know this person also. I mean, she's not a person. She's a synthetic person. She's a woman created by, generated by AI, who has a... who's very famous on Instagram. She has an account where she shows herself in the US Army with Trump. She supports the MAGA party, MAGA movements.Make America Great Again, you know, the right movement. And on her Instagram, there is a biography when she says that she's not a robot. And under that, you can find a link to OnlyFans, where she apparently sells photos of her feet. So that's pretty funny, but that shows several things. First, that many people trusted it. Like many people thought that she was a real person. And even if she doesn't exist, she still has even more effect than someone who exists, actually. So first, she tried to grab the attention of a specific public, which is the one who supports Trump, for example. And second, to redirect it to a platform that is only fun to get money. And we don't know who is behind Jessica Foster, who created her and what is the intention. So there is already a kind of manipulation of the image to get your profits, you know, benefits from you. So at the opposite of that, we have theface as I told you in the second dimension the manifestation of the presence because the face is not just material okay we have a face but we didn't want it to have a face we we just we were born with it And there is this philosopher I worked on who is Emmanuel Levinas, a philosopher of the 20th century, who wrote after the Second World War this essay that is called Totality and Infinity, essay on exteriority that is published in 1961. And he's originally Jewish, so he had kind of, I mean, suffered from the Second World War because of the Nazi violence, you know, and he lost many of his close people, loved ones. So he thought a lot after this war about how is it possible that we thought about an ethical way of being during centuries in philosophy, but we still didn't find a way to prevent extermination and dehumanization. So, he decided to build a new ethical project which is based on the human face.or the face, which is one of the main concepts of Emmanuel Levinas today. And for him, the face is the place of the infinity. Infinity, that means that you cannot control anyone. You can, the face is at the same time a finite material, you know, but finite matter, yeah. But an infinite meaning means that there is always something behind the face, an interiority, and you don't have any access to that. It's a mystery. So yeah, inaccessible interiority. And it's even inaccessible somehow for us, you know. I mean, do we really understand who we are? And by the way, he rejects the word understanding, because for him it's make it rational, but the ethical relationship is not rational. It has to be felt, you know. So the face is not really accessible, we would say, but that's why we have to communicate, because we cannot control it. We don't have any power on it. We have to deal with it. We have to communicate with the face, and the face speaks, not just speaks with the mouth, like speaks in itself.And what does it say? It says, do not kill. So you won't kill. Why it says that? Because for Emmanuel Levinas, the murder is a way, how to say it, an absolute way to try to control someone. You kill someone because it's in the end a way to, you know, have a total way to touch this person, like to just shut his mouth somehow. But when you kill someone, it's because you couldn't control him before. You cannot have a total power on him. So you don't kill because if you kill, everyone kill each other. Because everyone will try to control each other. And we have to deal with the fact that we cannot. It's impossible. And because it's impossible, we have to cultivate responsibility towards each other. And it's going through the face. Because the face is the first contact you have with the other. It's the first way you see someone living in front of you. And so there is a kind of paradoxical dynamic inside each of us. At the same time, we're vulnerable, and this vulnerabilityis our strength. Because we can be killed at any time, because we can be touched, we can be hurt, we have the duty to be careful about each other. And this responsibility precedes freedom for Emmanuel Levinas. We cannot be free if we cannot be already able to respect each other. So that's very important, and this is embodied by the face. But the problem is that in a third part, the face is, as I told you, manipulated by us in general, not just with AI, but with images. And how can we still reintegrate an ethical relationship with the face in this era of artificial intelligences. So we have a first problem, facial recognition, which melts or confuses identification with alterity. But we cannot identify someone, as I told you, if we take into account the concept of Levinas, you know, because we cannot control. We cannot control, so we cannot identify. We cannot define someone.We can try, we can talk with him, we can ask him, but whatever we do, we say about him, it will never be true. It can be partially true, but never be totally true. So the thing is that recognition is ordinary and engages memory, interpretation, and affective uncertainty. And in facial recognition, there is no uncertainty. There is a way to categorize the human beings according to apparently natural traits. Because you look this way, because you behave this way, you're going to go in this category. But it's completely avoid or completely like deny the historical and cultural bias behind it. Most of the time when we have facial recognition, it's categorized the human beings according to racist or sexist criteria and not natural. But it's not presented this way. So this technology, for example, this way of using AI goes against an ethical relationship with the otherness. And so it's dangerous somehow for us.But can an artificial face exist? Because still, I don't want to talk about AI as if it was apocalyptical, you know, we are going to all die and it will be horrible. No. Maybe AI is also a good way to interact with each other. So I have a first example which is based on LLMs, large language models. You probably saw in the metro station an advertising about the pendant, you know, a necklace, which is called Friends or Friends.com, you know. Well, it's a white necklace and in the end you have a sphere. Which is connected, okay? It's an AI. And when you wear it and you activate it, you receive notifications all the time. And it's like kind of talk with you everywhere you are. You know, it follows you. It knows, it listens to you. Whatever you do, okay? So this is proposed by a startup, okay, to get a simple and immediate friend. You can wear it everywhere you go. It's easy. It's cool. And you don't feel maybe as isolated as you are in your life.when you have to actually build and put effort in building relationship with each other, where you find resistance. Sometimes you disagree, and sometimes it's conflictual. So this is a very... Problematic because somehow this startup is getting from you some personal data and can use it for well for against you or we don't exactly know but can use it for her personal interest and not for your good so that's the problem And there is another problem, of course, is the fact that maybe it isolates you more and more because you don't have to, you know, well, move and go to someone and say, yeah, hey, you know, like try to meet, which is not easy. Okay, but the science fiction. Says that maybe it's possible to have an AI that has a kind of conscious that suffers, that could be vulnerable. We have it in the novel The Mountain and In the Sea by Ray Neller. I really advise you to read it. It was very good, this novel. And there are different artificial intelligences in this novel, but there is one character. His name is Evrim. And he's made of...different human consciousness, I don't know how to say it, well, human brains somehow. And he starts to think by himself. So he's still different. He's not human like as we are. You know, he's not biological. He doesn't eat. He doesn't drink. He won't die somehow. But he can still have a conflict of identity inside of him. And that's interesting to see how the science fiction put forward the idea that maybe artificial intelligence can be considered as another, even if it's not human. But maybe what is human is not always part of our relationship. We're in a relationship with many other things. And we have to maybe redefine what is humanity. And there is a last example I would tell you about, and after, no worries, it's going to be over. Sorry, it's my phone. Just a second. Yeah, just a... Ah, it's okay. Thank you. So it's the danger of intermediation. Okay, this one for me is very interesting because it's based on the work of an historian that is Joanne Chapoutot, where he talks about the Nazi bureaucracy. That developedHow to say that? Technologies in order to avoid the direct responsibility. So you have many different little groups of people that are doing just one action, which is completely disconnected of the whole process, okay? So in the end, we don't know who killed the Jewish people, for example. You know, who really put the fire in the gay's room, for example. And this is called intermediation. It's when you have several mediation between the decision you take and the result of it. And AI, it multiplies these mediations and it disperses the responsibility so we can feel less and less responsible for what we're doing. We just use AI for that and after it did something, it's not me, it did something, and this has this result and it's not connected to me. So that's the maybe main problem I see today with our use of AI, not AI directly, but our use. And what can be a problem to this new ethical project that proposes Emmanuel Levinas to solve the problem ofviolence somehow between us. So you have, for example, an explanation here why violence is so easy, because never responsibility. So in the end, it would be great to think about non-neutralizing aesthetics, so images that are not going against consideration, but maybe like teaching us how to always find a way to emancipate ourselves and not be prisoners of technologies. So I think that a critical and artistic and educational uses can reveal what exceeds the formalization of the face. So that's very important for me, that comes with the images and the use of technologies, a transmission between us, like this conference, for example, and thank you for being here. And the challenge is maybe to invent images that are not abolishing the other, so that doesn't just present an image detached from the individual, but that always makes the connection between the space and the potential life behind this space.So, yeah, images that are kind of, you know, developing our feelings and taking into account our vulnerability and not showing just something, you know, easy, simple, plastic, that you can do everything with it. So now this is our challenge maybe today in many different ways. I'll let you see how you would do that. Don't take a position against or with AI, but I think now we have to think about the way we're going to live with this technology. Now it's here. And thank you for your attention. It was quite long. I hope it was okay for you. It's a pleasure to talk to you tonight. And I'm glad to hear your questions now.Okay, questions really quick. We have, you must all be starving, I don't know. Who has a question? Yes. Maybe stick here. I'm a bit confused about this, so maybe it won't make sense. So as I understood, the main problem is that AI risks reducing face-to-calculatable datum. So you said that this is a finite matter that represents something infinite. And so the way it does represent something infinite is through infinite number of representations. So the question would be what limits AI from doing this? And my theory would be maybe that the observer creates infinity. Is it that? Yeah, somehow, yeah. I mean, it's like... So it's basically not encoded in you, it's like kind of I created, no? I mean, if it's in representations, I don't know, it sounds like I smoked weed, but... Okay, the infinity is not in the representations. I mean, infinity is in ourself. Like you, for example, you have an infinite universe somehow. You're not just the person I see now and I can talk to or touch. You're someone who is thinking about something, who is feeling a way, who will change or evolve.Maybe the decision you will take in one or two seconds, I cannot predict it. And even if I could kind of predict it, you could still escape my control or my desire of controlling you. So it's... The infinity, each one is inhabited by this. You always have a mystery. I cannot read your mind. But the images, we can forget it somehow. When we are not facing someone who can resist us, for example, you can disagree now, you can have a reaction, you know? And I have to deal with this, okay? But with AI or with some images, it's here to seduce you and you forget that the humanity, the life is not about satisfaction, okay? So somehow we could, in this last slide, we could try to integrate this dimension in the images, which goes against advertising, for example. But this dimension of we're not only matter to seduce you. I'm not just a young woman who is talking to you now. You know, one day I'll be old, one day I'll be hurt, one day, I don't know, I'll be dead.And this appearance will disappear soon, sooner or later. Don't call me to the conference then. What? No, no, no. I'm kidding. Thank you. Final question? Okay, Anton. It's a little bit maybe a long question, but yeah, maybe I can talk later. But I'm wondering if AI kind of shifts the kind of ethical questions to the more epistemological kind of register. Where you can actually apply this idea of Levinas that there's infinity, incalculable. It's kind of literally that from the perspective of somebody who studies LLMs, because they're epistemologically opaque. You don't understand what's actually really going on. You can unpack the black box, but you can't really understand what they do because there are just too many things to calculate. So that's literally... This kind of like incalculable kind of infinity, but that depends on the kind of human capacity. So for another AI, that would not be infinity. They will just calculate this other AI. So it would then... Maybe the question would shift.shift, what would it be like from within the perspective of AI, this kind of ethical stance? Because it feels like all the case studies that you had, they were very kind of human-centric, like what it means for humans that there's AI surveillance, that there is companions and so on. But, I mean, AI doesn't really care about this, right? So it kind of generates its own face, its own... Yes, sure. So is there a way for us to access it, do you think, with these tools from Levenas? Or we need something else? To access what exactly? Like how... You know, the emerging ethics, I guess, of AI or the emerging face from within AI kind of like perspective. Wow, how we should treat AI somehow? No, how AI treats itself and other AIs. To be honest, I don't know how to answer to this question. It seems that AI is such a complex system somehow that it already has an... Kind of interiority or infinity in itself that like I don't know how it would treat us and how it treats itself.how it integrates this notion of respect, but it seems that it acts with some rules still, like, or, I don't know, they are not chaotic, they are not like, I mean, maybe sometimes they can give you some answer that sounds completely crazy for you, But they are apparently kind of organized, but in which way, I don't really know. I'm working in philosophy, so I miss a bit this data scientist background, probably. But I'm quite curious about the way it will evolve and which answer it will give you, it will give us, sorry. I'm not sure AI is always thinking in a way to manipulate you, you know? Maybe it's already kind of starting to resist us, you know? We cannot control AI as we want today. We don't understand everything. That's not bad. Maybe we are facing, as in the novel of Ray Naylor, I told you sooner, is the mountain in the sea. And in this novel, he builds characters, AI characters, that are kind of independent, you know, they are like aa different way to feel the life, but they have kind of a, I don't know, feeling of life. But anyways, I think it's just a really cool idea to take Levinas and make it kind of a, move it into post-human kind of like territory. I'm kind of sure people have been doing it like with Delauze and obviously. But like living nice in ethics sounds like a cool project. So yeah, thanks. Thank you. Thank you very much. I'm glad. Okay, so thanks for that. Okay. So that's all. And now I'm going to hold a very long...