The 2013 Spike Jonze film “Her” imagined a not-too-distant future where digital voice assistants become super lifelike. The plot is set in motion from the moment a husky-voiced artificial intelligence assistant named Samantha says, “Hi there.” And then it becomes a love story between man and machine. That seemed crazy, maybe a little gross, when the film was released. But since then we’ve seen Siri, Alexa and Google Assistant evolve and spread to more devices. We engage them while walking around the house, opening the fridge or working out.
In the movie, there’s a moment when Samantha asks the main character a personal question: “You mind if I look through your hard drive?” After a long pause, Joaquin Phoenix’s character says, “OK.”
That hesitation caught the ear of Amy Webb, the founder of the Future Today Institute. I asked her to talk with me about the tech on display in the film “Her” and its implications. She said that scene shows how naive we can be about our digital assistants. The following is an edited transcript of our conversation.
Amy Webb: The interesting piece that doesn’t get addressed is that, of course, [today] we’ve invited all kinds of companies into our lives to do that very thing. It’s just that now, we’re mostly dealing in text and humanish-sounding voices, but that’s going to be changing pretty soon.
Jed Kim: What are the technical obstacles that exist to us having the world of “Her”?
Webb: There are some technical hurdles to overcome just in listening and understanding us. There’s a ton of work that’s already been done that clearly shows that we’re closer to this future than we may realize. There’s a field called behavioral biometrics that’s in the process of being born. This is where the spoken interfaces aren’t just recognizing who we are, like, “It’s Amy, and she’s using her computer,” but how and what we are in any given moment. For example, if we’re talking a little faster than normal, or a little slower than normal, or the pitch of our voice has gone up, or it’s gone down, that system could know, for example, whether we’re sick, if we’ve got a little cold, or some allergies, or maybe if we’re depressed, or if we’re going through a manic phase. That research is well underway.
Kim: This movie is probably the most optimistic that I’ve seen of films when it comes to artificial intelligence. Are we shifting our attitudes toward what the singularity represents?
The idea that our current digital addiction … might someday soon extend to spoken interfaces I think is a real possibility.Amy Webb
Webb: That’s so interesting to me that your read on this movie is optimistic. Mine is totally not. Mine is that I found it to be incredibly bleak and depressing. In one of the key parts of the movie, [it] has to do with social isolationism and a future in which we prefer to talk to machines than to each other. And in a lot of ways, the [operating system] in the movie, the AI agent Samantha, becomes a stand-in for some of the real people that had been in Theodore’s life. The idea that our current digital addiction, which tethers us to social media, might someday soon extend to spoken interfaces I think is a real possibility.
Kim: Do you think all this is going to create more addiction to tech?
Webb: Historically speaking, the things that humans find sticky, that we enjoy, are those things that reflect the world as we know it back to us. I think that’s an interesting prospect when we’re talking about smart spoken interfaces. What happens when we are in a situation in which there are smart AI agents that have voices modulated to tones that are most pleasing to us, that we feel very much like we’re in control of, but in fact, we have very little control over? These systems are making choices and decisions for us that sometimes we ourselves would not make. That is an interesting prospect that was teased out a little bit in the movie, but I think we’ve already started to see play out in real life.
Kim: At one point, the main character starts to have doubts and hides from his AI because it’s pointed out to him that she can’t really challenge him, like a human relationship. And the answer that then allows him to move forward with his relationship is essentially his AI telling him to just accept it. And that sounds like something you’d be railing against.
Webb: Already in the year 2019, we’re using a tremendous amount of technology in [the film]. Because there’s been a financial transaction involved, we feel like we own that technology. People feel like they own their cellphones or that they own their computers. We really don’t. We own the rights to use the devices, but we are umbilically tethered to one or more big tech companies who are in control of those ecosystems that are very much making a lot of the decisions for us without asking us what we think. If what we crave in the future is a humanistic relationship where we want to be challenged or we want to have arguments — whatever it is that we want — the OSs have been programmed to get us to click more, to get us to use them more. I think it’d be a tricky thing for a developer to figure out not just what would please us, but what would displease us in a way that would itself be pleasing.
Related links: more insight from Jed Kim
What’s new in voice-assistant technology? How about shoes you can tie using Siri? Melissa Locker wrote for Fast Company that Nike is developing a new shoe that features its FitAdapt technology. No longer will you have to bend over to tie your laces. You can tell Siri to tighten or loosen them for you. Yes, you can tell Siri to squeeze your feet. It’ll connect to your Apple Watch or device, but not to the last lingering shreds of your self-worth. Looks like the price tag will be more than $350. As someone who lives in a shoes-off house, this is somewhat tempting to me. What I’d really shell out for, though, is an option that whips my guests’ shoes off as well.
The voice assistant is a big competition space, and a new entrant into it will be the BBC. The Guardian reports that the U.K.-based broadcast giant is aiming to rival Alexa with an assistant it will call Beeb. The BBC wants more control over where voice assistants send users, like toward BBC content. In fact, you can’t listen to BBC radio through Alexa anymore. That happened because Amazon refused to share listenership data.
My favorite part of this story is that the BBC is asking its staff to record their voices to help the software understand regional accents. Which reminds me of a clip from the TV show “W1A,” which is from, you guessed it, the BBC. Guys, it can’t get any more perfect. See for yourself.