[This pre­sen­ta­tion was post­ed in four pieces. Section mark­ers appear at the breaks as there were some­times brief gaps between them.]

Fox Harrell: Firstly, thank you Kenny and the orga­niz­ers for invit­ing me and hav­ing me here. And one of things I’ll talk about is that social activism and using tech­nol­o­gy for empow­er­ment, it’s part­ly out­reach and the activism with­in the world, but also becom­ing learn­ers, doers, cre­ators, builders of our own tech­nolo­gies. And also under­stand­ing the ways that the tech­nolo­gies them­selves build oppres­sive struc­tures. And so we’ll see a bit of that.

So here’s what I do. I direct the Imagination, Computation, and Expression Lab at MIT. And we research and devel­op arti­fi­cial intel­li­gence and cog­ni­tive science-based com­put­ing sys­tems for cre­ative expres­sion, cul­tur­al analy­sis and social change.

So today’s talk has two dif­fer­ent parts to it. So I’ll talk about com­pu­ta­tion­al iden­ti­ty tech­nolo­gies (That’s self imagination—how we imag­ine our­selves through social net­works, online vir­tu­al worlds, through games.), and the project I’m run­ning, which is National Science Foundation sup­port­ed, called the Advanced Identity Representation Project, or the AIR Project. And then an exam­ple ICE Lab project for social change that worked with the Truth and Reconciliation Commission of Liberia.

So this is my moti­va­tion, to start off with. In the real world we can cre­ative­ly rep­re­sent our­selves in dynam­ic ways. So, we can vary our ges­ture, our dis­course, our pos­ture, our fash­ion, life sto­ries, the way we tell our sto­ries. And all of this is with an astound­ing sen­si­tiv­i­ty to social context.

Now, com­put­er tech­nolo­gies like com­put­er games, social net­work­ing, and vir­tu­al worlds are much more prim­i­tive than what we do in the real world. And so Kenny intro­duced me as the way of the future. Actually I think the ways that we’ve learned to nego­ti­ate the world—so the black expe­ri­ence, the ways we’ve had to adapt tragedies for sur­vival, all of that I think is much more advanced than what we have here in these tech­nolo­gies. Because these tech­nolo­gies require us to rep­re­sent our­selves through com­pu­ta­tion­al data struc­tures, through algo­rithms. So they have much less nuance than we have in the real world. 

So it rais­es a few ques­tions for us. So that’s, how can we serve the human need for self-expression and iden­ti­ty con­struc­tion using the com­put­er? How can we rep­re­sent our­selves in dynam­ic ways or even think about our pow­er rela­tion­ships with oth­er peo­ple in the world, social issues, and oppression?

Let me give you an exam­ple of what the prob­lem is. Do peo­ple know these games already? This is World of Warcraft here on the left, and this is Elder Scrolls IV: Oblivion on the right. So, in Elder Scrolls IV, let’s con­sid­er how it might stig­ma­tize users. So, the osten­si­bly African” char­ac­ters are called the Redguard are described in the essen­tial­ist stereo­type of a black ath­lete. So you read the man­u­al or the instruc­tions, they’re the most nat­u­ral­ly tal­ent­ed war­riors in Tamriel. The Redguards are also phys­i­cal­ly blessed with hardy con­sti­tu­tions, quick­ness of foot. So what hap­pens in game­play terms is it gives you bonus­es even­tu­al­ly to your run­ning and jump­ing abil­i­ty. [audi­ence laugth­er]

We can go a bit fur­ther, so let’s look under the hood of the game. What this is is a chart of all the default sta­tis­tics for a char­ac­ter. You’ll notice things that are a bit inter­est­ing here. Like if you’re an orc and you hap­pen to be female— One thing I should say is that most of these games say racial divi­sions for fic­ti­tious races like if you’re an orc or a troll or an elf. This one actu­al­ly has Norwegian peo­ples, the Redguard which are black peo­ple, the Bretons which are the French. So it saves those kind of racial changes that are nor­mal­ly fan­tas­tic races and applies them to what are osten­si­bly real races. 

So if you hap­pen to be a female orc you’ll see here by default you’re ten points more intel­li­gent than your male coun­ter­part. If you’re a human Breton, which is the French group that I men­tioned, you’re twen­ty points more intel­li­gent by default than your Norwegian or black coun­ter­parts. So the choice of race and gen­der with­in this game results in ability-based stereotypes.

I was inter­viewed for a site, Boing Boing, which is a blog, and crit­i­cized a num­ber of dif­fer­ent games for the lim­it­ed abil­i­ties to rep­re­sent our­selves, and an inter­est­ing thing hap­pened, which was that that was reblogged again by a gam­ing site. The orig­i­nal Boing Boing arti­cle was called some­thing like Professor Fox Harrell and His Chimerical Avatar” [Chimerical Avatars and Other Identity Experiments from Prof. Fox Harrell”], how these avatars that change based on what you do, your emo­tions, and the way that you inter­act with­in the world.

When it was reblogged on this site called Kotaku, they changed the title to Making Avatars That Aren’t White Dudes is Hard.” They also changed the goal of what I was try­ing to accom­plish, since they said, Fox Harrell wants to cre­ate avatars that look, well, like he does.” Maybe they meant maybe my social cat­e­go­ry, not just a kind of pure nar­cis­sism. Anyway, it spawned a series of incen­di­ary com­ments about it. 

So I pub­lished anoth­er arti­cle through that same blog and sug­gest­ed ten dif­fer­ent ways that we can improve our avatar representation. 

And maybe it had an impact, because Elder Scrolls V: Skyrim—some peo­ple might know that—this is the new default Redguard char­ac­ter. If you play the game, this is what the char­ac­ter looks like now. And the inter­est­ing thing, though, is that my critique—the arti­cle had a big pic­ture of my face along with an avatar. So it was a bit inter­est­ing. Maybe they respond­ed to the arti­cle. But the thing is, regard­less of how the char­ac­ter looks it still is going to be twen­ty points less intel­li­gent than your French counterpart. 

…num­ber of ways that they actu­al­ly could have improved it. Say, tie a life sto­ry into the way that you con­struct your char­ac­ters or expe­ri­ence would be a very sim­ple kind change. 

But we can go a lot deep­er than just look­ing at default stats. Let’s look at the under­ly­ing iden­ti­ty ele­ments that are bro­ken down into data struc­tures. This is from a game called Neverwinter Nights. And if you look under the hood, actu­al­ly, you have data struc­tures for race, even blood col­or by race—something almost like the one-drop rule. And so chang­ing race in this game actu­al­ly does­n’t change the appear­ance of your char­ac­ter. But many items have racially-based bonuses. 

Gender is inter­est­ing if you look under­neath the hood. Because when you play you just see the char­ac­ter, but in the data struc­ture there are actu­al­ly five dif­fer­ent gen­ders. You might think that it’s some­what more expan­sive than say a bina­ry gen­der rep­re­sen­ta­tion. Anyway, if you hap­pen to be male, both, oth­er, or none—which are four of the gen­der representations—you have a default male body type. So in fact despite these five dif­fer­ent [gen­der] types under the hood, you have just two bina­ry gen­ders with­in it.

So the point of all this is that these ele­ments are built into the very data struc­ture of the games. So just tak­ing up a char­ac­ter in a vir­tu­al world and say­ing, I’m going to look like some­body dif­fer­ent than I am,” it does­n’t do that much, actu­al­ly, in order to think about our­selves through a dif­fer­ent lens or think about our­selves as a new iden­ti­ty or to empow­er our­selves. In fact, a cer­tain type of oppres­sion or dis­crim­i­na­tion is built into the under­ly­ing struc­ture of the code. So it means it behooves us to build new tech­nolo­gies or think about becom­ing cre­ators or does or learn­ers and builders of technology.

In social net­work­ing there are prob­lems, too. So for exam­ple this is an old Facebook page, but opt­ing in or opt­ing out to an iden­ti­ty is just say, join­ing a group. And so you see in Facebook, if you want to be [some­what?] Native American, opt into the Native American group, with its stereo­typ­i­cal so-called noble sav­age metaphors here. It’s quite a sim­plis­tic mod­el of group membership.

I won’t go through all of these, but there are a num­ber dif­fer­ent prob­lems that we have, whether it’s com­put­er games, social net­work­ing, vir­tu­al worlds, any of our online accounts, any of these dig­i­tal rep­re­sen­ta­tions. And it’s the fact that our iden­ti­ties are reduced to sta­tis­tics. Social cat­e­gories are reduced to just graph­i­cal mod­els and skins, so just noth­ing but appear­ance. Character change isn’t dri­ven by any­thing like emo­tion or actu­al­ly meet­ing peo­ple or what you do, but rather com­bat, spa­tial explo­ration, and acquir­ing objects.

In social net­work­ing, we have a sim­plis­tic mod­el that join­ing a com­mu­ni­ty is just click a but­ton and you’re a mem­ber of the com­mu­ni­ty. You don’t actu­al­ly have to live, work, breathe with the community.

Virtual worlds, we have— So what about states like tran­si­tion­ing or becom­ing? I was actu­al­ly on a pan­el once with a trans­gen­der dig­i­tal media artist. That was some­one who was actu­al­ly in a state of tran­si­tion. I said what about becom­ing? Well that’s some­thing we all deal with as you become an expert, for exam­ple, from being a novice. We’re always in states of tran­si­tion, but usu­al­ly you’re always just some­thing fixed in dig­i­tal tech­nol­o­gy. Limited cul­tur­al diver­si­ty. So there are a lot of lim­i­ta­tions that we have that in cur­rent technologies. 

But you might ask what’s even the big prob­lem just because well, it’s just a game, right? It’s just a vir­tu­al world. Well, there’s research—some of my own empir­i­cal research, Jerry Bailenson at Stanford is anoth­er person—that shows that changes to the vir­tu­al rep­re­sen­ta­tions impact issues like inter­per­son­al con­fi­dence, body image, stu­dents’ per­cep­tion of them­selves as learn­ers and does. So our real-world inter­ac­tion with peo­ple can be changed sig­nif­i­cant­ly by the way we inter­act through our vir­tu­al iden­ti­ties. And if you even have an account for your cell phone, you have a vir­tu­al iden­ti­ty in some kind of way. It’s not just these pyrotech­nic exam­ples in the video games I mention. 

So, address­ing these prob­lems I think will make more diverse user groups, pro­vide bet­ter cus­tomiz­abil­i­ty, make for more salient and pow­er­ful expe­ri­ences, invent new forms of art, enter­tain­ment, and iden­ti­ty. So in short, we can do a lot bet­ter than the cur­rent state. 

So you can say that we’ve at least been try­ing to take some mod­est steps towards doing so in the Imagination, Computation, and Expression Lab. And so with the sup­port of the National Science Foundation I ini­ti­at­ed a five-year project that’s called the Advanced Identity Representation Project. That’s devel­op­ing a toolk­it that works across platforms—social net­work­ing, avatars, pro­files, characters—that enables rich self-expression; that can change dynam­i­cal­ly; address issues like social stig­ma, bias, prej­u­dice; and is based on cog­ni­tive sci­ence mod­els of how we cat­e­go­rize in the world.

So instead of just the naïve intu­ition of a game design­er who might not know any­thing about the expe­ri­ence of diverse groups of peo­ple, we’re look­ing at how we actu­al­ly cog­ni­tive­ly cat­e­go­rize in the world. And it’s much dif­fer­ent than peo­ple assume. We don’t actu­al­ly just have cat­e­gories [where] we just try to force peo­ple into small box­es. What peo­ple tend to do is to have a pro­to­type, mod­els of fam­i­ly resem­blances. It’s much more flex­i­ble than the way that peo­ple imag­ine in most of these infrastructures.

And we can learn a lot, even going back to 1903 from say WEB Du Bois and double-consciousness. So as I men­tioned, before the kind of rich ways that we nav­i­gate the world give us a lot of pow­er and insight in how we can devel­op more pow­er­ful tech­nolo­gies. So the dual aware­ness of peo­ple from mar­gin­al­ized groups of their self-perception and social stig­ma that tends to be attrib­uted to the group.

Everyday self-presentation. So the fact that we adapt to per­form iden­ti­ties like ges­tures and dis­course, reg­is­ter for dif­fer­ent social situations.

And even recent­ly in 99, iden­ti­ty torque. That’s the psy­cho­log­i­cal­ly painful expe­ri­ence of a per­son­’s self-conception dif­fer­ing from broad­er stig­ma­tiz­ing per­cep­tions rein­forced by clas­si­fi­ca­tion infrastructures.

So I mean it’s just essen­tial­ly the same thing that Du Bois said back in the day, but just the fact that these infra­struc­ture we have, like the game for con­struct­ing our char­ac­ters, or online accounts, actu­al­ly result in psy­cho­log­i­cal­ly painful expe­ri­ences where you have some dis­junc­tion between who you are in the real world and then the way that you have to rep­re­sent your­self through those technologies. 

Audience 1: Do iden­ti­ty torque again. Just explain a lit­tle more.

Harrell: Sure. So, it’s essen­tial­ly what Du Bois was talk­ing about, but the idea of torque is the idea of twist­ing. It’s a twist­ing of our biogra­phies, of our life sto­ries, against the tech­nolo­gies that we have to engage. So when you have to go and use— It’s that sense I have say cre­at­ing the char­ac­ter in Skyrim, and oh, it looks like me now, but I have to be a lit­tle bit less intel­li­gent if I want to look like me. So maybe I should look like the French char­ac­ter or Roman char­ac­ter. So my biog­ra­phy being twist­ed up against that infrastructure. 

And the same thing when you cre­ate an online pro­file on Facebook and you have to kow­tow to com­mer­cial inter­ests, for exam­ple. Or you have to rep­re­sent your­self in a group just by opt­ing into the group. So it’s that sense that you have much rich­er self and sto­ry, and when you go to use these tech­nolo­gies then there’s a kind of psy­cho­log­i­cal­ly painful expe­ri­ence that’s detri­men­tal to your health and happiness.

So, the sort of things that our toolk­it can allow are say chang­ing user’s self-representation for dif­fer­ent social groups. Basic things that peo­ple do—code-switching. So you have dif­fer­ent expe­ri­ences. Again, the things that we do in the real world can give us a lot of insights for new technologies.

For bet­ter or for worse, right? So, stereo­typ­ing or pass­ing. Are peo­ple pass­ing on Facebook? That does­n’t just mean racial pass­ing, but say what if we’re— I think Kenny and I were talk­ing just ear­li­er about a move­ment towards say punk rock music or skate­board­ing with­in the black com­mu­ni­ty. What if peo­ple just want to seem a bit more like they’re in that scene than they actu­al­ly are? There’s a kind of pass­ing there. So can we iden­ti­fy this kind of pass­ing that we have with­in social networking? 

What about swap­ping between mul­ti­ple iden­ti­ties the way that we use dif­fer­ent enti­ties in dif­fer­ent situations? 

What about aspi­ra­tional iden­ti­ties? Not just a fic­ti­tious iden­ti­ty or some­thing neg­a­tive like pass­ing, but what about what you see your­self as in the future? We mod­el our aspi­ra­tional identities.

And one of the things that I’ve noticed is that in social net­work­ing or com­put­er games, at some lev­el there’s some lev­el of abstrac­tion or some sim­i­lar­i­ty between the way we rep­re­sent our­selves in these dif­fer­ent tech­nolo­gies. And so what I did here was just draw a graph that shows a Facebook pro­file and a char­ac­ter in a game called Dragon Age. And so in Facebook you have your pro­file, you have these links like friends with” and your groups of friends. You have dif­fer­ent pages, dif­fer­ent things that you like there. And in Dragon Age you have dif­fer­ent skills and class­es like you might be a rogue, or you might use a dag­ger. So there’s some lev­el at which you can abstract and start to com­pare these iden­ti­ties to process them in the same way. You can start to say is this iden­ti­ty like anoth­er iden­ti­ty? And so that’s where you can start to use com­put­ing to think about these kind of issues, like Du Bois and oth­ers have thought about.

…go much into all of the detail, but I’ll tell you about some of the kind of things our sys­tem allows. So let’s say dis­cov­er­ing social cat­e­gories. You might look at every­body in your social net­work and then find who are the peo­ple that tend to like the same thing? The peo­ple that like cer­tain films, what are the kind of movies that they tend to like also. Or what are the kind of activ­i­ties, places that they go. And so you can find social cat­e­gories that are built into the net­work, that peo­ple haven’t explic­it­ly gone in there and labeled. 

What about find­ing peo­ple that are like you, not just by going in and say­ing I want to find every­one who’s a woman who lives in Cambridge,” but to say what about the things that you do or the things that maybe you like?

You can pre­dict belong­ing in cat­e­gories and say, based on what I know about these oth­er social cat­e­gories, I think that this per­son would fit into this oth­er cat­e­go­ry.” And some of these aren’t just util­i­tar­i­an tools. They’re also just tools to think with, to think about issues. Like how can you think about belong­ing from a com­pu­ta­tion­al per­spec­tive? Because most com­put­er sci­en­tists don’t deal with issues like belong­ing as a mem­ber of a group.

Seeing iden­ti­ties in terms of one anoth­er. So that’s again this kind of pass­ing, or say even pos­ing. Is a per­son try­ing to be a mem­ber of my group? And so you could see their iden­ti­ty through the lens of a group that you find your­self in.

I’ll show you just a lit­tle bit of what this looks like. I’ll scroll through it. You can do things, say see the sys­tem with a few dif­fer­ent games like say a game, StarCraft, Civilization here, and then from there say what do oth­er peo­ple in this per­son­’s social net­work who like these games also tend to like?” And so this is an MIT stu­den­t’s social net­work. And for his net­work for some rea­son, these gamers also tend to like some series of these clas­si­cal music com­posers. That’s some­thing that you would­n’t have known that ahead of time, before using the system.

You can go in and rela­bel the cat­e­gories, too. This is a cat­e­go­ry called hip­sters,” and maybe [inaudi­ble] inde­pen­dent films or oth­er sort of things. Then you find things in oth­er cat­e­gories that the per­son might just tend to like. Or things in peo­ple’s par­tic­u­lar local com­mu­ni­ty, even. So, sports fan.

So the inter­est­ing thing, too, is it’s not just try­ing to cat­e­go­rize all sports fans, it’s very local for that per­son­’s net­work. So when you talked about punk rock music before, one per­son­’s punk rock music might be pop punk rock, or Green Day, or some kind of band. Someone else might be under­ground. Someone else might be strict­ly black punk rock. It could be inter­na­tion­al. So it’s very par­tic­u­lar for your local com­mu­ni­ty. So what the sys­tem can do is find okay, who like sports in his own net­work and then what oth­er things they like. They tend to like the Cold Stone Creamery. But also they tend to like some spe­cif­ic restau­rant that just hap­pens to be in Watsonville, California. So it’s not just any­thing, it’s some­thing very par­tic­u­lar to their own network.

And so now we can go back to those things things I talked about before, iden­ti­ty torque, double-consciousness, and pass­ing, things we know about from the real world, and say— We actu­al­ly imple­ment­ed those in a com­pu­ta­tion­al way. So when I talk about pass­ing, we could say— The mech­a­nism you could use to think about pass­ing on a com­put­ers is alter­ing some of your rep­re­sen­ta­tion to more close­ly resem­ble a mem­ber of anoth­er cat­e­go­ry. So you have all of these kind of char­ac­ter­is­tics cod­ed in the data struc­ture, and to think about it in ways in social net­work­ing to anony­mous­ly explore iden­ti­ties that are dif­fer­ent than your own, whether for good or for ill. So it’s just a way we can say that these phe­nom­e­na exist but in tech­ni­cal terms that are also in dia­logue with what we do in every­day life. The same with double-consciousness.

Multiple visu­al rep­re­sen­ta­tions. We built sys­tems, for exam­ple a game where depend­ing on what you do—if you’re walk­ing through the sub­urbs in your char­ac­ter, then if you act aggres­sive­ly or if you pray or if you punch some­one, the char­ac­ter changes in a total­ly dif­fer­ent way. If you use your cell phone or take out your wal­let, you start to become more commerce-oriented, like a tycoon. Like a Monopoly man, you have stock charts burst­ing out of your head, or mon­ey bags start to appear. So your char­ac­ter’s just con­stant­ly chang­ing. It’s not just get­ting new weapons, you’re chang­ing in more poet­ic ways. 

What I want to do now is just segue to anoth­er project which is a quite dif­fer­ent but relat­ed to iden­ti­ty dis­cus­sion, because it helps us to tell sto­ries from the point of view of dif­fer­ent enti­ties. And this project is called the Living Liberia Fabric. We’ve imple­ment­ed a num­ber of dif­fer­ent projects. This one was ini­ti­at­ed [by] a col­league in inter­na­tion­al affairs, but it’s quite time­ly now, espe­cial­ly since Ellen Johnson Sirleaf was just award­ed the Nobel Peace Prize. 

Many of you might know that the West African nation of Liberia was nev­er for­mal­ly col­o­nized. It was polit­i­cal­ly estab­lished in 1847 by free­born African Americans, Africans freed from cap­tured slave ships, all of whom were set­tled in an area that would become Liberia by the American Colonization Society. So this was a col­lu­sion of inter­ests. It was slave own­ers, the US gov­ern­ment… So I’m not say­ing that it’s a kind of benev­o­lent expe­ri­ence. In fact, the con­flict between the local pop­u­la­tions with a series of a prof­i­teers who who aid­ed them ini­ti­at­ed civ­il wars from 1989 until 2003. About 250,000 peo­ple were killed. That’s one third of the pop­u­la­tion that was displaced.

And so one of the things that Ellen Johnson Sirleaf did was to appoint a Truth and Reconciliation Commission mod­eled on the more famous one in Post-Apartheid South Africa. And one of the things that the TRC said is that memo­ri­al­iza­tion is a nec­es­sary part of fur­ther­ing the peace process. And so that’s—

And so I had my stu­dents meet with the Truth and Reconciliation Commissioner from Liberia, peace muse­um experts, dias­po­ra Liberians who were in Georgia, also sur­vivors of civ­il war—a gen­tle­men whose own son was abduct­ed to become a child sol­dier, tragically—and went around and col­lect­ed a num­ber of sto­ries and pub­lished books about this. So we real­ly engaged with the com­mu­ni­ty, spent a long time just to try to under­stand that con­flict. So see­ing our­selves as stake­hold­ers, too. Not like we’re going to use tech­nol­o­gy for out­reach but say­ing how do we relate to this through our own his­to­ries, our own biographies.

And so what the sys­tem does is it uses users’ actions to reveal mul­ti­me­dia con­tent, video, pho­tographs, texts, and an AI sys­tem I wrote called GRIOT that simul­ta­ne­ous­ly coor­di­nates themes, nar­ra­tive struc­ture, and media assets, by find­ing analo­gies between them.

And so what we did was find a series of infor­ma­tion through our expe­ri­ences talk­ing to peo­ple, we encod­ed it here at three dif­fer­ent levels—a visu­al lev­el, at the lev­el of the kind of frame that it inter­acts with, and then stake­hold­ers groups, dif­fer­ent kinds of themes, activists themes. And so we entered all this infor­ma­tion in a form that could be manip­u­lat­ed com­pu­ta­tion­al­ly. You’ll have to read this. It’s fas­ci­nat­ing, the for­mat that our sys­tem could use. And then behind that, it struc­tures a nar­ra­tive dif­fer­ent­ly each time you inter­act with it.

So what I’ll do is I’ll show you just a lit­tle bit of the way that the sys­tem works. And so the text is a bit a bit hard to read on the pro­jec­tor. Here it says a silent moment for loss.” So what you do, before I start it, is you have a series of dif­fer­ent fig­ures here. When you start it, you just hear a bit of ocean sound. If you click on one of the fig­ures, what that does is pick a stake­hold­er group, like woman or child. But that stake­hold­er group could mean woman sur­vivor, it could be woman com­bat­ant. So you’re not choos­ing every­thing, you’re choos­ing one aspect of it.

That will pick a series of clips that will appear with­in the fab­ric. And then from those clips, you can click on just one of them. Then the next clip will be sim­i­lar to the one you just picked. So if you click on some­thing about a woman that deals with activism, then the one that comes after that will deal with activism. Then the next one might be a child that deals with activism. So it’s a way that we’re using AI not just to gen­er­ate a sto­ry but rather to impro­vi­sa­tion­al­ly make sure that there’s the­mat­ic links between it. So I’ll just play a lit­tle bit of it so you can see what what it looks like.

[The demon­stra­tion is view­able in the orig­i­nal video from ~2:305:25. There is also a demon­stra­tion clip on YouTube, show­ing a dif­fer­ent instance of the project.]

So here you see all the oth­er relat­ed stake­hold­er groups, and also kind of lost peo­ple that might be relat­ed to that group.

I should also say it’s a bit arti­fi­cial show­ing it here. It says some­thing like this is what the tech­nol­o­gy can do. Because the idea of the sys­tem was to say not what can we do for Liberia but to say how can we root com­pu­ta­tion with­in the cul­ture of Liberia. So that’s when you can ground com­put­ing prac­tices with­in diverse cul­tur­al prac­tices. And so it’s a diverse per­spec­tive. We’re not just say­ing how can we go there and cul­tur­al­ly plun­der for the sake of tech­nol­o­gy, and so it’s a big dif­fer­ence here.

And these videos were col­lect­ed by a col­league actu­al­ly in Liberia, dri­ving around in a truck from doc­u­men­tary film, from archival footage. So there’s a close per­son­al con­nec­tion to a lot footage, too.

So I’ll wrap up here before mov­ing on to Malia. And so I’ll just say that the con­clu­sion is just that com­put­ing can be used for sub­jec­tive expres­sion and social change. It’s not just a kind of objec­tive medi­um. And so in the kind of work we do, we think about imag­i­na­tion. Imagination as a kind of art­ful thought, inter­ac­tive nar­ra­tive, poet­ry, gam­ing, as a space for self-expression and self-imagination. Computational iden­ti­ty sys­tems, imag­i­na­tion as cat­e­go­riza­tion, and more robust, more pow­er­ful forms of cat­e­go­riza­tion that learns from life expe­ri­ence, learns from the expe­ri­ence of the mar­gin­al­ized or the sur­vivors, and the dig­ni­ty of peo­ple who strug­gle for social change. And our projects are just a few mod­est steps towards those ends. Thank you.