Cynthia Breazeal: I think there are count­less amaz­ing oppor­tu­ni­ties for arti­fi­cial intel­li­gence and its impact on soci­ety. I think one of the areas I’m tru­ly the most excit­ed about is edu­ca­tion. We know that there’s tremen­dous inequity in qual­i­ty of edu­ca­tion world­wide. We know that there are over a bil­lion peo­ple who can’t even read at an ade­quate lev­el, and with the Internet spread­ing they’re going to have lim­it­ed abil­i­ty to access and lever­age the ben­e­fits of the Internet if they can’t read. And with AI there’s the poten­tial for deeply per­son­al­ized learn­ing expe­ri­ences for peo­ple of all ages and stages—even for work­force retrain­ing, right. 

Challenges around that always go to issues of pri­va­cy. So if you’re going to try to cre­ate AIs that deeply under­stand us in order to per­son­al­ize and adapt and opti­mize to us, it’s going to have to learn about us. And of course that is this kind of dove­tail edge to the sword between what should it be learn­ing and adapt­ing to you to ben­e­fit you, and how does it pro­tect your pri­va­cy and your secu­ri­ty so that you’re not put at risk in unin­tend­ed ways? There are many diverse voic­es that need to be a part of that con­ver­sa­tion.

In the case of edu­ca­tion, I have a project right now where we’re lever­ag­ing AI and tech­nol­o­gy to bring lit­er­a­cy to the world’s most under-resourced com­mu­ni­ties, inten­tion­al­ly. And there the stake­hold­ers are the chil­dren, the teach­ers, the par­ents. Building a net­work of sci­en­tists and con­tent providers who want to cre­ate tech­nolo­gies that help these pop­u­la­tions. And even­tu­al­ly, gov­ern­ments. Anyone who would even­tu­al­ly become a stake­hold­er of that kind of tech­nol­o­gy. But we are inten­tion­al­ly build­ing a net­work of peo­ple from all those per­spec­tives in order to do that.

We real­ly want to show that AI is not just for those few who can mas­ter this tech­nol­o­gy, but we’re real­ly going to apply this with a human­i­tar­i­an val­ue sys­tem, to say we want to under­stand how to design this tech­nol­o­gy, want to under­stand from work­ing with those stake­hold­ers across all these dif­fer­ent areas how to best inte­grate it. We’re going to learn from that process. Because you’re always going to learn from the process of inte­gra­tion. There’s very few cas­es where you can drop a tech­nol­o­gy into a work­ing human sys­tem and have it just work. It takes time to iter­ate and devel­op, and iter­ate, and learn and learn, and then you final­ly come to solu­tions that work.

And if we’re cre­at­ing AI for human beings, we’d appre­ci­ate the expanse of human experience—how we think, how we learn, how we make deci­sions, what mat­ters to us, to design tech­nol­o­gy to sup­port those val­ues.


Help Support Open Transcripts

If you found this useful or interesting, please consider supporting the project monthly at Patreon or once via Square Cash, or even just sharing the link. Thanks.