Why I’m not afraid of Sophia

An excellent article in Fast Company, “My three decades at Disney taught me not to fear automation”, by Duncan Wardle, reminded me that I had wanted to write about “meeting” Sophia, a robot created by Hanson Robotics. Back in April, I agreed to give a lecture at Talent Land in Guadalajara (Jalisco, Mexico), on the same stage and immediately before Sophia was due to speak. Like Duncan, I too was curious as to whether people like me would at some point be replaced by robots. Furthermore, I had specifically asked to speak at the event next to Sophia, I wanted to make sure I had a moment before or after my conference to get to “meet” her.

After my 45-minute talk ended, I was introduced to Sophia just before she was taken on stage. I say “take” because, although she now has legs, back then she had her badge but no means to move around, and was on a platform covered with a cloth, which was then wheeled on to the stage. From there, she was connected via a long cable to a laptop with a carer at the controls who told me that the Sophia we were to hear was one of four copies: one with permanent residence at the headquarters of Hanson Robotics in Hong Kong and the other three roving around the world giving interviews and lectures.

My first impression was that this was an impressive piece of interface development: Sophia is not — at least, not at the moment — a conversational robot or endowed with any kind of artificial intelligence: she’s an experiment in conversational interface, with a texture skin eerily similar to that of a person, who maintains her robotic appearance by means of a transparent parietal and occipital that reveals some mechanisms inside her head and hands, which are printed in 3D in white plastic and is endowed with very lifelike expressions… but only reproduces sound files and combinations of movements issued by her carer via a computer, to which she is permanently connected. Interviews are prepared in advance, with answers already prepared to previously scheduled questions.

Hanson Robotics is also working on the development of a certain conversational intelligence, and Sophia’s gestures and expressions are impressive: to see her put on an expression of surprise, condescension, intrigue or irony by raising an eyebrow is a worthwhile experience and one that certainly makes you think.

At the same time, I have to admit I found myself looking round to see if Sophia had attracted a bigger crowd than me, or if the audience was paying more attention to her presentation. I’m pleased to report that the central hall of the pavilion was equally full during both conferences, and I think I can say that people were paying as much attention to me as Sophia, and what’s more, she didn’t do a questions and answers session. That said, I wouldn’t be either surprised or offended had it been the other way round, after all, it’s not every day you get the chance to attend a presentation by a robot.

Which is all well and good, but what would really happen if, besides generating greater expectation, Sophia delivered a better lecture than mine? Could robots like Sophia replace that part of my work which consists of simply transmitting ideas to an audience? I make a point of involving the audience as much as possible in my lectures, and I think that Sophia, at least in her current state, would not have been able to deal with questions or other interaction, or to lead a discussion. It’s possible, with the right preparation and audiovisual support that she could match or even do better than me. But even so, it wouldn’t be Sophia per se giving a lecture, but a person or a team, using Sophia as a prop. To be honest, I think it would be marvelous if Sophia or a similar robot prepared a conference on her own; that said, conferences make a significant contribution to my income, so if robots started takin over, maybe I would rethink my position and contact Hanson Robotics or whoever and try to work with them, contributing my two decades of experience as a lecturer to developing Sophia or whichever robot, while thinking how I could continue contributing value in this new scenario.

Can a robot give a better lecture than a human? Absolutely. Can it do so in an interactive way, generating different conversational strategies depending on the participation of the audience and connecting their comments with the points being covered? Not yet, but perhaps in some distant future. Any concerns I might have are far outweighed by my desire to see technology progress, and in response to such developments I would be focused on developing more interactive conferences, focused on providing a different learning experience, on getting people to think in new ways.

Like Duncan Wardle, I’m not afraid of automation: I have long believed that it will improve how I work in the future. And my meeting with Sophia has simply reaffirmed that belief.

(En espa?ol, aquí)