# Recommendations Algorithms, Freedom, Genhumanism Approved speech wording and attribution are preserved; paragraph formatting is awaiting review. Visual text and AI-generated descriptions remain separate from speech. Paragraph placement is editorial; recording display times and exact frame timestamps are unchanged. ## Visual — Locked room *DISC-0033::VIS-001::OCC-01 · Display occurrence: approximately 0.000–248.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 100.000000 seconds in original video.* **Visible slide text — not speech** You’re inside a pleasurable room that you don’t want to leave. Silently, someone locks the door from the outside. Are you free? **AI-generated visual description — not speech or approved wording** An open, sunlit room contains a suspended hammock. No image credit is printed. --- ### Other speaker (context) speech **Other speaker (context):** Thank you so much, James. **James Brusseau:** Great. Well, thank you for that kind introduction. And thanks for joining us today. There it is. Yes. So we were just talking about freedom out at the cafe. We should actually move ourselves out to the cafe today. It took us about 30 seconds to discover that the entire field is so complex, blurry and knotted that we could almost say nothing. We're both just overwhelmed by how vast and difficult questions about human freedom are. What is freedom? How do we know we have it? How can we get more of it? How do we measure it? No answers to those questions. However, what we plan to do today is at least make some progress into this mess of uncertainty along, at least in the beginning in my presentation, along a single line. That is, along the line of recommendation algorithms. Think just simple, like the Netflix movie that you see that pops up on your screen after you're done watching one thing suggests something else. I'm going to talk a little bit first about why I think recommendation algorithms are important and why they present an interesting route into this question of freedom and servitude. Then in the second part, I'm going to talk a little bit about a specific strategy that one might be able to employ with respect to recommendation algorithms to enhance a certain version of human freedom. And then in the third part, I should have put my little clock here, but you stopped me. But in the third part, we will talk briefly about an idea that Carlo has been promoting and discussing this idea of the right of discontinuity and attach that to the idea of gen-humanism and what that means. So that's the roadmap, at least of the first presentation. And the first presentation's first slide, you've had a good chance to look at it. It's a hard question this one. If you are in this room where you have everything you want, all your needs are satisfied, your wants, your desires, you have no desire to leave, you're not going to leave, you don't want to leave, but you can't. Are you free? Now, this dilemma comes from the English philosopher John Locke in 1685. His book was The Nature of Human Understanding, something like that. And here's what he said. He said that we should distinguish between the freedom to get what you want, which you have in this room, and the freedom to change what you want. And for him, this idea of being able to change what you want is fundamental to what it means to be human. And then, as a footnote, if we had more time and this were a philosophy class, then we would go from that step to say, look, when we change what we want, we are also changing who we are as individuals. But we're not going to go there. Instead, we're just going to start with this idea that one way of thinking about freedom is having the ability to create new preferences for ourselves, to explore new and different interests. And that's what I'm trying to sum it up with this slide in this dilemma. --- ## Visual — Recommendations and locked rooms *DISC-0033::VIS-002::OCC-01 · Display occurrence: approximately 250.000–386.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 300.000000 seconds in original video.* **Visible slide text — not speech** Recommendation algorithms and locked rooms... Zero moment: Captured by established preferences SHETLAND SEASON 1 EPISODE 1 Starts in 0 seconds **AI-generated visual description — not speech or approved wording** A television character fills the background; an autoplay tile for Shetland appears at lower right. No further image attribution is printed. --- ### James Brusseau speech Now, next, how does that dilemma connect with recommendation algorithms? Well, I'm just watching my own screen here. Yes. Yeah, so there is this moment when you're watching Netflix, where the one show ends and then right away, while the credits of the one show ending are going, the next show is already beginning. And I think that that experience, if it's done right, when Netflix does what it wants to do, that is when Netflix hits you with this new movie just at the moment where you're about to ask, what should I do next? Or what do I want to watch next? Just when you're going to ask that question, bang, they get you with the next show. And so what they're doing is not just that they're giving you another series or movie, which is like the one you saw before. It's that they're cutting off and eliminating this ability to ask yourself, well, do I want to have a different interest? Do I want to do something else now? And that's not necessarily a bad thing. Usually it's a good thing. Usually I'm really happy about that. Or just in front of my TV, I just want to be satisfied. I don't want to be free. I just want to be happy, content, some sense blissful in my ignorance of my possibilities. Nevertheless, I think that this question about the locked room that John Locke asked, I think it is thick in this particular human experience. I think we are giving away something palpable, some real human freedom and we can define what that freedom is. It's the ability to change what you want or to create new preferences. We're giving that away to get the comfort of the next show. And then just as a quick footnote, it seems to me that just as a graphic for it, your freedom is being cut off here, according to my argument. --- ## Visual — Diabolical restriction *DISC-0033::VIS-003::OCC-01 · Display occurrence: approximately 388.000–464.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 420.000000 seconds in original video.* **Visible slide text — not speech** Diabolical: Not restricted by opposition of others, but by similarity with our own past **AI-generated visual description — not speech or approved wording** A carved face has an hourglass superimposed over its mouth. --- ### James Brusseau speech But as opposed to the usual perception we have of that experience, this is not something that someone else is doing to me. It's not like someone is taking away my right to do something. It's not really Netflix that is doing this to me. The operative force, what is taking away my own freedom, is my own interests, my own preferences. It's the fact that I have told Netflix that I like, I can't show you, but I like detective programs with kind of vaguely European overtones. And I watch those over and over again. So the fact that Netflix then gets me with a different detective program with European overtones, that's not Netflix. It's me doing that to myself. So it's a strange, this freedom is in a strange dynamic for it's a struggle of me against myself instead of me against other people. --- ## Visual — Beyond film recommendations *DISC-0033::VIS-004::OCC-01 · Display occurrence: approximately 466.000–838.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 600.000000 seconds in original video.* **Visible slide text — not speech** Film recommendations are insignificant part of life, but... - Netflix - Twitter/X (Social media echo chambers) - LinkedIn - Tinder *So many options, need recommendations... DANGER **AI-generated visual description — not speech or approved wording** A yellow danger sign shows a person slipping. --- ### James Brusseau speech Okay, so that's the first thing I want to say. Look, there's this idea of human freedom. It connects with recommendation algorithms. And then I admit that Netflix is trivial. Who cares about movies? But I think that this same structure of uncertainty about our own decisions, I think this is rapidly expanding to include more aspects of our life and it's rapidly penetrating more deeply into our lives. First, on the idea that it is expanding to encompass more regions of our life. We see, for example, in social media, it could be Twitter or Instagram or Facebook, it doesn't matter what, this phenomena of echo chambers or filter bubbles where everyone gets channeled into their own tribal group. So if you like a certain kind of music, then you keep getting posts about that certain kind of music. You're not exposed to anything different. If you have a certain political stance, you keep getting posts that reinforce that political stance. You don't see anything different. This is something that has been talked about widely on the level of society. I think that's just a repeat of the same structure that we discussed for human individuals on the social level. And then expanding in a different direction, there's the, to me at least, the very serious question about what kind of profession you will pursue. What kind of job you will have. Well, more and more people find their jobs through LinkedIn and similar platforms. And how do those platforms promote or present jobs to you? The same way Netflix presents movies. They isolate your established interests and preferences, and then they give you more of the same. So LinkedIn is very effective for most of us, most of the time at presenting opportunities, maybe for jobs or maybe in our field to give a presentation, that sort of thing. It does that very well, but silently and in the background of that activity, the same kind of freedom diminishment is happening. That is, we're losing the ability to choose or try different jobs. So, I mean, just as a simple example, I mean, I remember when I was your age, one thing I really loved to do is go to a different city and get a job as a waiter. Because that way you meet the people, interact with the scene, it's easy. It's easy to work and so on. Now, as it happens, as a philosopher, not surprisingly, I was a terrible waiter. Usually I would stay in a place until I was fired from my job. Even in Europe where it's harder to fire people, they would still fire me somehow. But it didn't matter back then, and I was relatively good at giving lectures and doing philosophy. That was something I knew but no one else did. Today, that's not true. Today, I could get a job as a philosopher at one or another place because people would see that on LinkedIn I have a certain kind of success. But at the same time, at least if we had LinkedIn when I was young, if I wanted to do, and sometimes I do, if I wanted to do what I did when I was young and just say, you know what, forget these books, forget this thinking about freedom and that stuff. I want to go work as a waiter. I could not do that anymore because the same machine, LinkedIn, that tells everyone I'm decent at philosophy also tells everyone that he's got a bad, bad history as a waiter. He keeps getting fired. I have two months here, one month there, that kind of thing. So LinkedIn for you guys, it's trapping you in the trajectory that you are setting up. It's a good trap. I mean, you probably want that, but it's also taking away to a certain extent from your ability to try different things. No one's going to give you a job outside of your training because they know you're not suited for it. That didn't affect, I mean, we're different. We come from a different place. We could just pick up our suitcase and disappear to a different city and be whatever we wanted to be. That's not true for you guys. So Netflix, Twitter, LinkedIn, and then Maximum Case. I think it's Maximum Case with whom you form a family who you love. That sort of thing. Tinder or Hinge or whatever choose your platform. Same thing, recommendation algorithms. They are going to keep giving you, you know better than I do, the same man and the same woman with a different name over and over and over again. That's what they do because that's what works. They can figure out what works by seeing how many conversations you exchange, how long the relationship goes on and so on. And if you disappear from the platform, well, that means it really worked. Then when you come back, they're going to find someone else who's just like that, but again, with a different name. Again here, I mean, that's not a bad thing. I mean, I've been through the wars, the romantic wars long enough to tell you. Yeah, you kind of do keep coming back to the same thing. But still, there is a limitation that these platforms impose upon you in terms of the children you will have, the life you will have fundamentally. --- ## Visual — How couples meet *DISC-0033::VIS-005::OCC-01 · Display occurrence: approximately 840.000–880.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 860.000000 seconds in original video.* **Visible slide text — not speech** HOW COUPLES MEET IN THE US Recommenders increasingly determine human experience: *Who we are...* Chart labels: Online; Through friends; Through family; Work; Bar; Grade school; Neighbors; College. Horizontal ticks: 1950, 1960, 1970, 1980, 1990, 2000, 2010, 2020. Vertical ticks: 0%, 10%, 20%, 30%, 40%, 50%. Source: "How Couples Meet and Stay Together": a longitudinal study of social life in the US by M. J. Rosenfeld, Reuben J. Thomas, and Sonia Hausen **AI-generated visual description — not speech or approved wording** The red Online curve rises steeply after the 1990s and ends above 50%; the other displayed meeting routes decline toward 2020. This describes the drawn curves; no point values have been inferred. --- ### James Brusseau speech Okay, so this – just a statistic. This to me is really quite striking, that you can see. So, you can see when I was your age, most people met through friends, but in the last 10 years, that's more than 60%, at least in the US, of existing couples met online. To me, that's staggering. If that's true, then recommendation algorithms really are controlling what future we will have as families, the way we will compose ourselves at that level. Okay. So now I'm going to wrap up this first part, right? What I'm trying to do is I'm trying to say there's questions about freedom and there's a definition of freedom, which is freedom means the ability to change who you are, to develop new interests and preferences. --- ## Visual — Problems: individuals *DISC-0033::VIS-006::OCC-01 · Display occurrence: approximately 882.000–934.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 900.000000 seconds in original video.* **Visible slide text — not speech** PROBLEMS... For individuals: an ethical problem because human freedom is confined by our established preferences. Our ability to change is restricted by our own past. [Film cover] BILL MURRAY · ANDIE MACDOWELL GROUNDHOG DAY SPECIAL 15TH ANNIVERSARY EDITION EXPERIENCE HIGH DEFINITION **AI-generated visual description — not speech or approved wording** The individual-freedom argument is paired with the Groundhog Day film cover. --- ### James Brusseau speech Recommendation algorithms tend to give us what we want, which is a different definition of freedom, But they do tend to quell or diminish the ability to change what we want. And this leads to three kinds of problems. The first problem is the main one I've been talking about. It's an ethical problem, which is that we are trapped in our own past. We are unable to escape the men and women we have dated, the jobs we have had, the training that we have built. --- ## Visual — Problems: society build *DISC-0033::VIS-007::OCC-01 · Display occurrence: approximately 936.000–938.000 seconds in the original video. Boundaries are estimated.* *Short-display inclusion exception:* Short substantive build, reference, or contact content; retained explicitly despite normal five-second threshold. No durations aggregated.  *Exact frame PTS: 936.000000 seconds in original video.* **Visible slide text — not speech** PROBLEMS... For individuals: an ethical problem because human freedom is confined by our established preferences. Our ability to change is restricted by our own past. For society: a solidarity problem because users are funneled into echo chambers of narrow perspectives and polarization. Diversity collapses into tribalism. **AI-generated visual description — not speech or approved wording** A second argument is added beneath the individual problem. This short intermediate build is retained because it adds substantive content. ## Visual — Problems: platforms build *DISC-0033::VIS-008::OCC-01 · Display occurrence: approximately 940.000–946.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 940.000000 seconds in original video.* **Visible slide text — not speech** PROBLEMS... For individuals: an ethical problem because human freedom is confined by our established preferences. Our ability to change is restricted by our own past. For society: a solidarity problem because users are funneled into echo chambers of narrow perspectives and polarization. Diversity collapses into tribalism. For platforms: an economic problem because bored users leave the site. **AI-generated visual description — not speech or approved wording** Three bent arrows introduce the individual, society, and platform problems. A photograph of bundled cash accompanies the platform argument. ## Visual — Solution *DISC-0033::VIS-009::OCC-01 · Display occurrence: approximately 948.000–1042.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 980.000000 seconds in original video.* **Visible slide text — not speech** SOLUTION Recommendation algorithms re-engineered to catalyze new curiosities, interests, and preferences. --- ### James Brusseau speech We can't get outside of it. Then there's also a social problem, which is this idea of the filter bubble and the same mechanism put onto the social level. And then we're moving past here. There is also as it happens, the platforms also have this problem, by the way, that people do get bored on Netflix. People do leave on Netflix because they don't get a good variety of shows. So the platforms are themselves aware of this problem and just as a footnote actually I just won a small grant from the some board or something to actually try to develop recommendation algorithms for companies like Netflix to help them help users discover different kinds of movies different kinds of interests to help them get outside of what their preferences would tell you they want. We're going to talk about that next, but let me say that to close the first part, we've got a definition of freedom, which is the ability to change your interest. We've got a set of problems that arise from that ethical, social, and business level. And then we have a solution, which is, okay, if this is the problem and this is so terrible, you say, Professor Rousseau, then what should we do about that? Well, the answer is to re-engineer our recommendation algorithms so that they can catalyze new curiosities and interests instead of just repeating the ones we've already had. Okay? Okay. Okay. So, that was part one. All right? In part two – how long have I been? **Speaker uncertain:** No, no, no. Okay. **James Brusseau:** Okay, so in part two now what I'm going to try and do is talk a little bit about usually my audience includes computer scientists so typically I talk a little bit about technical solutions to philosophical problems and that's what I'm going to do here but it will not be too technical I promise so we have the problem which is the recommendation algorithms keep you trapped in your preferences. Let me show you how Netflix does this, and everyone does it the same way. They have two strategies for recommending your next movie. One is called content filtering, and the other, the more important one, is called collaborative filtering. --- ## Visual — How recommenders trap preferences *DISC-0033::VIS-010::OCC-01 · Display occurrence: approximately 1084.000–1098.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 1090.000000 seconds in original video.* **Visible slide text — not speech** HOW RECOMMENDERS TRAP ME WITH MY OWN PREFERENCES: NETFLIX **Content-based filtering** Similar movies Liked by Bob Recommended to Bob **Collaborative filtering** Liked by Alan and Bob Similar users Liked by Alan, recommended to Bob **AI-generated visual description — not speech or approved wording** Left: a user’s liked movie links to a similar movie recommended back to that user. Right: two users share liked films, and another film liked by Alan is recommended to Bob. Yellow highlighting emphasizes similarity. ## Visual — Content-based filtering example *DISC-0033::VIS-011::OCC-01 · Display occurrence: approximately 1100.000–1156.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 1120.000000 seconds in original video.* **Visible slide text — not speech** Content-based filtering [Left poster] HARRISON FORD IS THE FUGITIVE [Right poster] HARRISON FORD È BLADE RUNNER **AI-generated visual description — not speech or approved wording** The Fugitive and Blade Runner posters are displayed side by side. The repeated actor name is visibly prominent. *Visual limitations:* Small Fugitive production credits and the smallest Blade Runner footer are unreadable at source resolution; not reconstructed. --- ### James Brusseau speech Content filtering is just like this. If you liked Harrison Ford, The Fugitive, well, they will look for other movies that are like that. And they define a resembling movie because it has Harrison Ford. Again, Harrison Ford the Fugitive is called Action Adventure. Harrison Ford and Blade Runner is also called Action Adventure. Just that is enough to say, well, if you liked The Fugitive, then you get Blade Runner. Now what makes that logic work, and I'm going to keep coming back to this, what makes that logic work is the fundamental idea. And here the philosophers can talk a little bit. The fundamental idea that similarity is what rules logic. Everything should always try to be the same. That's what we're looking for. Make things the same, the same kind of movie. --- ## Visual — Collaborative filtering example *DISC-0033::VIS-012::OCC-01 · Display occurrence: approximately 1158.000–1222.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 1180.000000 seconds in original video.* **Visible slide text — not speech** Collaborative filtering [Left user] HARRISON FORD IS THE FUGITIVE BILL MURRAY · ANDIE MACDOWELL GROUNDHOG DAY PARAMOUNT PICTURES PRESENT MARLON BRANDO Al Pacino · James Caan · Richard Castellano Robert Duvall · The Godfather [Right user] HARRISON FORD IS THE FUGITIVE BILL MURRAY · ANDIE MACDOWELL GROUNDHOG DAY ? **AI-generated visual description — not speech or approved wording** The left user has three films, including The Godfather; the right user has the first two and a question mark. *Visual limitations:* Some small poster credits are unreadable; repeated poster text appears here as displayed content. --- ### James Brusseau speech Now that same logic, this is the more important way that Netflix works. They try to find two people who are similar right here, both philosophers, lovers, both about the same age. One Italian, one who's presumably French. I guess that makes us different, actually, right? About the same age, right? And then they look and they see, oh, both he and I—of course I'm making this up, I don't pretend to know his movies—he and I like The Fugitive. He and I like the movie Groundhog Day. He likes The Godfather. So what are they going to recommend to me? It's the same idea. Similarity is driving the way recommendation algorithms work. And that's why you're getting caught in the preferences that already exist, and that explains the fact that you keep getting on Tinder or on Facebook or whatever you're looking at, why you keep getting the same thing over and over again because resemblance rules their logic. --- ## Visual — Antagonistic filtering *DISC-0033::VIS-013::OCC-01 · Display occurrence: approximately 1224.000–1276.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 1250.000000 seconds in original video.* **Visible slide text — not speech** ANTAGONISTIC FILTERING ESCAPE ESTABLISHED PREFERENCES CREATE NEW PREFERENCES Liked by Axel and Alice Dissimilar users Tangential similarities Liked by Axel, recommended to Alice **AI-generated visual description — not speech or approved wording** A prohibition symbol crosses the shared-films route. Dissimilar user figures connect through tangential similarities; a film liked by Axel is recommended to Alice. Yellow highlighting emphasizes creating new preferences. --- ### James Brusseau speech Now, what I am promoting is not collaborative filtering, which was the previous one, but antagonistic filtering. What I want to do is get some computer scientists together, and maybe there's space for lawyers in this, not sure, get some computer scientists together and see if we can find a way to take people who are not similar but dissimilar, man and woman, young and old, U.S., Italy, or U.S., Japan, and so on and see if we can start from that universe, the set of people who are different, distinct, and then find within that universe between some few of these people a little bit of an overlap. So something like this. So here are two people who like really different movies. I'm not sure if anyone has seen Napoleon Dynamite. This is a highly recommended, very idiosyncratic movie. As a quick footnote, Napoleon Dynamite is very difficult for the recommendation algorithms because it really does appeal, just statistically you can see this, to an unpredictable person. --- ## Visual — Difference before similarity *DISC-0033::VIS-014::OCC-01 · Display occurrence: approximately 1278.000–1350.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 1300.000000 seconds in original video.* **Visible slide text — not speech** ANTAGONISTIC FILTERING: DIFFERENCE BEFORE SIMILARITY MY NAME IS SALT Napoleon Dynamite BRING HOME THE YEAR’S SWEETEST COMEDY! A murdered wife. A one-armed man. An obsessed detective. The chase begins. HARRISON FORD IS THE FUGITIVE ? - Different age - Different gender - Different nationality - Different language - Similar taste in restaurants - Different preferences in music - Different interests in pastimes **AI-generated visual description — not speech or approved wording** Different film tastes are arranged above two dissimilar user figures. “Similar taste in restaurants” is highlighted among the listed differences. *Visual limitations:* Small poster reviews, festival laurels, and production credits are partly unreadable. No missing attribution is inferred. --- ### James Brusseau speech They have no way of knowing who will like that movie and who doesn't. It causes great trouble for Netflix. But anyway, so one person likes Napoleon Dynamite, one person likes My Name is Salt, another person likes Harrison Fugitive, and they're completely different people. But maybe you find that they have similar interests in restaurants, or maybe you find they have similar interests in music or something. So you use the power of big data or AI, because AI stuff, you use the power of AI to both locate differences and then to find after you have the differences, find a few telling similarities, and then use that to recommend one movie to another. And the idea is, the purpose, circled in yellow here, the purpose of this is to help people open up new kinds of interests. It's not just for people to find a Harrison Ford movie that they did not know existed. --- ## Visual — Process *DISC-0033::VIS-015::OCC-01 · Display occurrence: approximately 1352.000–1416.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 1380.000000 seconds in original video.* **Visible slide text — not speech** PROCESS | START WITH: DIFFERENCES | THEN SEEK: SIMILARITY | THEREFORE: RECOMMEND | |---|---|---| | Distinct preferences in movies between users; * Difference: Age; * Difference: Language | Shared favorite restaurants and/or music and/or pastime and/or profession ... | One person’s favorite movie recommended to other. Purpose: Open new range or genre of interests | **AI-generated visual description — not speech or approved wording** Arrows connect differences → similarity → recommendation; the purpose statement is circled in yellow. --- ### James Brusseau speech No, it's to help people find movies that are in a genre that they would not have thought of looking for themselves. It's to help people find a kind of movie which is completely outside of what the individual would have found were they looking on their own. So the idea of antagonistic filtering is the idea of helping people discover new interests and preferences, and that traces back to the idea of serving human freedom, which is my main interest. And if you have a social interest in working against tribalism and filter bubbles, then the same logic would apply. --- ## Visual — Opposites attract *DISC-0033::VIS-016::OCC-01 · Display occurrence: approximately 1418.000–1462.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 1440.000000 seconds in original video.* **Visible slide text — not speech** INTUITION: OPPOSITES ATTRACT (SOMETIMES) **AI-generated visual description — not speech or approved wording** A film still pairs a woman with a ribbon in her hair and a man in a black jacket. No printed image credit is visible. ## Visual — Provocation and curiosity *DISC-0033::VIS-017::OCC-01 · Display occurrence: approximately 1464.000–1614.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 1500.000000 seconds in original video.* **Visible slide text — not speech** CONCEPT: PROVOCATION IS BETTER THAN ACCURACY Same but different *or* Different but the same = Curiosity **AI-generated visual description — not speech or approved wording** One painted portrait branches by two arrows into two differently styled portraits. The images illustrate variation around a face. No full artwork attribution is printed; a small signature appears on the lower-right work. --- ### James Brusseau speech Okay, so just very quickly, the intuition behind this idea is opposites attract. I don't know if that's true or not. If this were a less formal setting, I would ask what you think of that. The general answer I get is people say that opposites attract. It works for a while, but it doesn't work for very long. That's what people say. I'm not sure. My belief is that opposites really do attract and it is possible for people to develop new kinds of interests but if you don't think that then everything I'm saying is not going to hold water because basically you do have to be open to things you do have to have some kind of curiosity that pulls you out of the experiences to which you are accustomed and then the last thing to do or almost the last thing I think just checking came out. No, no, no. Yeah, just for myself. Right. So the last thing now to do on this section is to say, so I think that a kind of opposites attract or curiosity is very important for this. And so I tried to think a little bit about what you can, like, what is curiosity? When do we have it? How do we get it? And I think that the curiosity fundamentally comes from this idea of a logic of difference. Like you find some kind of way of being engaged with something which is fundamentally different from what you would expect, which is fundamentally dissimilar with your own experiences. It's those two things. It has to be completely different from what you would expect, but there must be some kind of engagement. And I think that it's captured a little bit by these three pieces of art. So that's your own Caravaggio Joe as the woman. And then to the Van Gogh, I mean, I think that's actually a move of similarity, right? You kind of say, well, it's a different way of rendering the same woman. But when you go to Picasso, no, I don't think so. I mean, I don't think that's a move of similarity. Like when you look at that picture of, I guess, a woman, I mean, you don't think it's a woman, right? You think it's shapes, something. And it takes you a while, or at least it takes me a while They kind of make the connection. Oh, I see. This actually connects with the Caravaggio. It's just different ways of imagining a woman in a painting. So I think something like that is at the root of curiosity and I think something like that is what is functioning, I think, in this notion of antagonistic filtering that I just described. And then I think something like this would be the key to having the kind of freedom that I'm interested in, which is the one, which is the ability to choose, create new interests and preferences. Okay? Okay. So just very quickly, and this is the end of part two now, you could take that same idea and move it through Netflix, Twitter, LinkedIn, Tinder. I mean, all these, you could use the same logic for everything. --- ## Visual — Dissemination *DISC-0033::VIS-018::OCC-01 · Display occurrence: approximately 1616.000–1632.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 1620.000000 seconds in original video.* **Visible slide text — not speech** DISSEMINATION - Netflix - Twitter/X (Social media echo chambers) - LinkedIn - Tinder .... ## Visual — Conclusion *DISC-0033::VIS-019::OCC-01 · Display occurrence: approximately 1634.000–1662.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 1650.000000 seconds in original video.* **Visible slide text — not speech** CONCLUSION *Recommendation algorithms can do more than give us what we want, they can also help us to want new things* - Catalyze human freedom - Reduce polarization/tribalism - Increase platform profits **AI-generated visual description — not speech or approved wording** Three bent arrows introduce the listed outcomes. --- ### James Brusseau speech You can use it to find new interests on social media, find new jobs, to find new men and women in your life. And then the conclusion of part two is if you accept my arguments – maybe you don't – but if you accept my arguments, then this will help us – this idea will help us to catalyze human freedom. It will reduce polarization and tribalism in society. And it will also make money for Netflix, which actually – so I'm not French. I'm American. So that's always hovering somewhere in there. All right. Okay. So that's part two. Part three is – so I should kind of wrap this up now, right? Very quickly. --- ## Visual — Right to discontinuity *DISC-0033::VIS-020::OCC-01 · Display occurrence: approximately 1678.000–1806.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 1700.000000 seconds in original video.* **Visible slide text — not speech** POSTSCRIPT: RIGHT TO DISCONTINUITY, GENHUMANISM **Right to discontinuity** Right to be forgotten Right to escape one’s own PII Right to privacy as *inalienable* *Kim Kardashian selling her PII... **AI-generated visual description — not speech or approved wording** A composite face is assembled from vertical strips of different faces. --- ### James Brusseau speech Just two ideas for part three. One is this idea of the right to discontinuity. As I understand it, Professor Casanato has been discussing this a bit with you. To me, at least, I will just talk quickly because you have an expert right here. I'll just say quickly, it seems to me that this connects a little bit with the right to be forgotten and maybe also with a philosophical idea that some rights are inalienable, by which we mean freedom is one of those. According to John Locke, you cannot sell yourself into slavery and you cannot freely get rid of your freedom. It's impossible on the level of the being of freedom itself. It can't make sense for you to freely give away your freedom. And maybe, I'm not sure, but maybe personal identifying information is something like that. So maybe, even though Kim Kardashian, let's say, and makes a lot of money selling away her personal information. We know everything about her more than we want to know. Maybe you could make the claim that even though she has sold that information to a TV channel or whatever, maybe she has a right to reclaim that information even if it was sold because people have an inalienable right to their own personal information. So just like in philosophy, we say, that people cannot freely sell themselves into slavery maybe you could say something similar about privacy and the right to discontinuity because the right to discontinuity as I understand it is this ability to escape from your personal information make some new stuff so you could buy, give back your personal information, destroy it, delete it, become someone new. --- ## Visual — Genhumanism *DISC-0033::VIS-021::OCC-01 · Display occurrence: approximately 1808.000–2116.000 seconds in the original video. Boundaries are estimated.*  *Exact frame PTS: 1900.000000 seconds in original video.* **Visible slide text — not speech** POSTSCRIPT: RIGHT TO DISCONTINUITY, GENHUMANISM **Genhumanism** Production of new personal identities through time = production of new interests and experiences Not evolutionary but emergent Not escape from old PII, but overwrite it [Background sign, partly covered] CZECH Restaurant Pub **AI-generated visual description — not speech or approved wording** A black-and-white portrait in historical dress is placed beside a modern color portrait outside a pub. The slide prints no names identifying the pictured people. --- ### James Brusseau speech But that's not my area. My area is instead a gen humanism. So this is the last slide. This is the end. I have two stories of gen humanism for you that I think serve fairly well to summarize what I have been talking about today. One is the story I've kind of already partially told, the story of Isabel Eberhardt, The woman in black and white at the close of the 19th century, she was a minor aristocrat in Switzerland, got tired of her existence, escaped to North Africa, and became an avatar of today's reality. She became a man, though I don't think she did it quite the way we understand that. But she dressed as a man, adopted the role of a man in Arab society, adopted the local religion, adopted the local diet. Her project was a project in becoming someone different. It's a radical example of how someone created new interests, how they generated new preferences for themselves across the entire span of their existence, in terms of the language they spoke, in terms of the clothes they wore, in terms of the food they ate, in terms of the god they worshipped. In all these ways she changed what it meant to be her. So gen-humanism is the idea that it is a part of human nature to make ourselves anew, to create new identities for our existences. That opposes the more traditional philosophical belief, which is that the core of our nature is to perfect ourselves and be the best version of ourselves in one way or another. So I'm substituting the idea of change for perfection. I'd rather be different than be perfect. That's the idea of genuinenism. That's Isabel Eberhardt. And then the last example is contemporary, a guy named Max Hawkins. He developed a curious tool for Facebook. What his tool does is it fed him – and he was a computer scientist despite appearances – it fed him random events. He was in San Francisco, so he would get invitations to random events, things that were outside of his established interests in Facebook. So here he is. He was invited to, or there was a Facebook announcement for a gathering at a Czech pub where people of Czechoslovakian descent were going to gather and drink and talk about Czechoslovakia. He was not Czechoslovakian, not a big drinker, but the invitation came, so he went. This is a way, obviously, of developing new interests and preferences. You throw yourself into a different kind of reality. And so he did that for various in many different ways. And I think it's an instantiation of the kind of human freedom that I've been discussing so far. What I do want to say about his strategy, however, is I think it's imperfect. That is, he got his suggestions through random slots on Facebook. It was just a random number generator that shows a post and then sent it to his wall to read. What I want to do is do better than random. I want to not – you can always, for example, that's what Netflix does too. Netflix does want to open you up to new things. So once in a while, they will hit you with just a random thing. Tinder does it also. You know, some crazy – really? Her That happens. But what I'm trying to do is imagine a kind of way of algorithmically creating recommendations which are, again, outside of the scope of your preferences and interests and your past, but which are nevertheless engaging, interesting, attractive. So I think Max Hawkins gets half the way there, but there's still work to be done. Okay, so that is a discussion about what freedom is, one definition of freedom. There are many definitions of freedom. Then the discussion of how that definition of freedom fits into AI reality today, recommendation algorithms. Then how it crosses through an area of law, the right to discontinuity, where your professor is studying. And then finally, how it intersects with what one guy is doing with his knowledge and computer science today, and so that's the direction that we will start off with for our presentation. --- ## Visual — Curiosity Engine project contact *DISC-0033::VIS-022::OCC-01 · Display occurrence: approximately 2118.000–2122.000 seconds in the original video. Boundaries are estimated.* *Short-display inclusion exception:* Short substantive build, reference, or contact content; retained explicitly despite normal five-second threshold. No durations aggregated.  *Exact frame PTS: 2120.000000 seconds in original video.* **Visible slide text — not speech** Curiosity Engine The Curiosity Engine is an NSF funded project in AI Contact: James Brusseau, Pace University NYC, jbrusseau@pace.edu **AI-generated visual description — not speech or approved wording** A QR code accompanies the project and contact statement. It is preserved in the image; its destination has not been inferred.