In a world of con­flict­ing val­ues, it’s going to be dif­fi­cult to devel­op val­ues for arti­fi­cial intel­li­gence that are not the low­est com­mon denom­i­na­tor. In Asia par­tic­u­lar­ly, we have a lot of coun­tries that believe that gov­ern­ments are the best way to make deci­sions for peo­ple, and that indi­vid­ual human rights can be voiced only by the state. In that con­text, if we are try­ing to come up with arti­fi­cial intel­li­gence that rec­og­nizes indi­vid­ual human rights, and that looks to empow­er cit­i­zens and user, and to cre­ate trans­paren­cy, it’s going to be real­ly chal­leng­ing to come up with an inter­na­tion­al coor­di­nat­ed regime that does this.

People have devel­oped AI that is pre­dic­tive. People are research­ing ways to make sure that AI is able to tar­get adver­tise­ment at peo­ple depend­ing on their pref­er­ences, the devices they use, the routes they take. Now, that kind of pre­dic­tive AI can very eas­i­ly be used for sur­veil­lance. And it’s a fact in Asia that the states in Asia, includ­ing India, are invest­ing very heav­i­ly in mass sur­veil­lance and they’re cre­at­ing large cen­tral­ized data­bas­es that they haven’t ful­ly worked out how to sweep as yet.

In India we’ve also got news of the state using drones to mon­i­tor assem­blies of peo­ple in pub­lic places just to make sure that noth­ing goes wrong. We’ve got news that the gov­ern­men­t’s devel­op­ing social media labs that are sup­posed to watch the online social media to see what peo­ple are say­ing and what kinds of sub­jects are trend­ing. And in that con­text, again, the ques­tion that we’re ask­ing our­selves is, when the state choos­es to use its resources to get AI to do these things, how far is AI going to be used to con­trol and mon­i­tor the cit­i­zen as opposed to enabling the cit­i­zen. Because in democ­ra­cies like ours, the bal­ance of pow­er between the cit­i­zen and the state is real­ly del­i­cate, and there is a great poten­tial for AI to tip that bal­ance of pow­er in favor of the state.

While it’s impor­tant to make sure that we don’t chill inno­va­tion, it’s also impor­tant to be cau­tious and to make sure that tech­nol­o­gy does­n’t drag us down a dark path. We’ve got exam­ples from his­to­ry like the Manhattan Project, like the way in which tech­nol­o­gy was used dur­ing the Holocaust, to remind us that if we’re not care­ful about what we do with tech­nol­o­gy, it can be abused in ways that that we will come to deeply regret. So it’s nec­es­sary to make sure that we have human rights, polit­i­cal the­o­ry, but also all the oth­er dis­ci­plines that under­stand what it means to be human and how to engage with humans involved in the design­ing of AI.

If we don’t work out a way in which cit­i­zens are able to ask the right ques­tions about AI to ensure account­abil­i­ty every time AI is cre­at­ed and used, we might be head­ing towards the world that Orwell pre­dict­ed, and that would be real­ly unfor­tu­nate because new tech­nol­o­gy should lead to a bet­ter world and not a more con­trolled world, or an unequal world.

As you know, tech­nol­o­gy is eas­i­ly sold to peo­ple, and it moves very quick­ly around the world. And so it’s real­ly impor­tant to inter­vene in Asia at the stage of design. People some­times have the best of inten­tions, but because of the way in which they’re edu­cat­ed, or the way in which they’re taught to think, the way in which they design tech­nol­o­gy can end up being real­ly dam­ag­ing to the world. Conversely, it could end up being real­ly beau­ti­ful as well, and that’s why it’s real­ly impor­tant that we get into AI right now and help the peo­ple that are design­ing it think of it in a way that imag­ines a bet­ter world.