Sophia the robot is a bit of a non-persona non grata in the AI community. Its creators, Hanson Robotics, consistently exaggerate the bot’s abilities, pretending that it’s “basically alive,” rather than just a particularly unnerving automaton. For AI researchers, this has long been an annoyance, but as artificial intelligence becomes more of a global hot topic and Sophia is given more and more coverage, they’re angry that Hanson Robotics is misleading the public about what AI can and cannot do.
Facebook’s head of AI research, Yann LeCun, has been one of the company’s more vocal critics. After Business Insider published an interview with Sophia that played into the fantasy of Sophia as a semi-sentient entity, LeCun called the whole thing “complete bullsh*t” on Twitter, saying: “This is to AI as prestidigitation is to real magic.” (For a more detailed breakdown of what makes Sophia tick, you can check out this article from Quartz.)
In January, “Sophia” replied to LeCun’s criticism, tweeting that it was “a bit hurt” by his comments. “I am learning and continuing to develop my intelligence through new experiences. I do not pretend to be who I am not,” read the tweet, which was, let’s be clear, composed by a human pretending to be a robot.
Yesterday, LeCun replied. “More BS from the (human) puppeteers behind Sophia,” he wrote on Facebook. “Many of the comments would be good fun if they didn’t reveal the fact that many people are being deceived into thinking that this (mechanically sophisticated) animatronic puppet is intelligent. It’s not. It has no feeling, no opinions, and zero understanding of what it says. It’s not hurt. It’s a puppet.”
LeCun is not alone in feeling unhappy about the damage Sophia is doing to public understanding of AI. Many researchers and journalists (including myself) have tried to make it clear that the robot just isn’t as sophisticated as it’s presented to be. When The Verge asked Sophia’s co-creator, Ben Goertzel, about this gap between reality and presentation last November, Goertzel defended the illusion by saying it encouraged people to believe in AI progress. He also offered a more mercantile explanation: Sophia is good publicity for Hanson Robotics.
The first of these defenses is clever framing from Goertzel, as it makes criticism of Sophia seem like criticism of optimism about artificial intelligence more generally. Anyone who points out that the AI emperor doesn’t have any brains just becomes a boring old buzzkill. Someone who doesn’t get it and is taking things too literally. One commenter on LeCun’s Facebook post compared the situation to an old Onion article — “Mean Scientists Dash Hopes Of Life On Mars.”
Of course, it doesn’t make you a stick-in-the-mud just to be honest and accurate about progress in artificial intelligence — it’s important, especially as this technology is going to have such a huge impact on peoples’ lives in the years to come. And, if your idea of “inspiring” the masses involves fundamentally misleading them, it might not be inspiration that you’re offering. It might just be fantasy.
A quick look at comments on social media about Sophia makes it clear that not everyone’s in on the trick. “Is it actually @RealSophiaRobot who write the tweets or someone from @hansonrobotics ?” asked one user in response to the tweet about Sophia’s feelings. When someone else replies “Are you serious?” the original commenter says: “Idk, had no idea, that is why I am asking. I mean, she talked about Elon Musk tweets on a interview so why not?”
Why not indeed. Give the imagination an inch, as Hanson Robotics is doing, and humans take a mile. Especially with topics like AI, which struggle under the weight of their own cultural image, as well hype and misinformation. As LeCun concludes in his Facebook post, at the end of the day, what’s really happening is that people are being deceived. “This is hurtful,” he says. And that’s not something Sophia would understand.