Could Watson and Siri love like Her's Samantha?
Source: Jessi Hempel
Samantha, the hot and sexy operating system in Spike Jonze's Oscar-nominated film Her, has a lot in common with the IBM (IBM) computer system Watson, the bold and brainy 2011 winner of the nationally televised television show Jeopardy! They both read quickly. They both process natural language, understanding with relative ease the difference between a tiger in the wild and Tiger Woods. They both learn on their own, self-programming an ever-deepening knowledge base. But unlike Samantha, Watson has yet to fall in love.
For those who have so far missed Her, a surprisingly believable sci-fi rom-com, a recap: At some point far enough in the future that high-waisted pants are in for men (but not so distant that kitchen appliances have substantially changed), a man falls in love with his computer operating system. Theodore Twombly, played by Joaquin Phoenix, is an ordinary guy in the midst of a divorce who buys a smartphone-size device loaded with a program advertised as a "consciousness." Like all new technology products, it is intended to be one thing, but left to its users, it becomes another. In Theo's case, it boots up and then promptly names itself Samantha. He has a hand in this, of course. He chooses Samantha's gender and offers early decisions that will inform her personality. From there, however, Samantha develops on her own, combining Theo's cues with all the other data she is rapidly ingesting from the world.
Her is an epic love story, successful because it skillfully convinces an audience of the possibility that Samantha has programmed herself into the ability to have and experience feelings like a flesh-and-blood human. We, the audience, are persuaded to champion the love affair we witness; we become convinced it's mutual.
Jonze's film comes closer than its antecedents -- from the 1968 Stanley Kubrick film 2001: A Space Odyssey to the 1980s television series Small Wonder -- to depicting technology not as it exists in the future, but as it is evolving right now. When the hit 1999 film The Matrix repackaged for the masses the idea of sentient machines, able to control the human experience of reality, most people couldn't yet imagine a reality in which machines took on human attributes. In a Windows 98 world, it was pure sci-fi.
In recent years, that has changed. Artificial intelligence has come to market in a number of mainstream products. Apple's (AAPL) frustratingly deficient personal assistant Siri, present on the company's mobile devices, mostly lets me down when I ask her personal questions. But even though I understand that she is artificial, for a moment when she responds to a question with humor -- "I'm old enough to be your assistant," she replies tartly when I ask how old she is -- I feel warmly toward her.
Similarly, while reporting early stories on IBM's Watson in 2010, I visited the lab where the supercomputer practiced his moves for years before his TV debut. I watched the expression on a programmer's face as he cheered on the swirling globe avatar, breath held. When Watson got an answer right, I thought I saw a glimpse of exuberance. We all slip readily into anthropomorphizing our current technology, and that gets easier as natural language processing gets better and as our computers, quite literally, get smarter.
In fact, we encourage anthropomorphizing. As MIT Technology Review points out, studies are testing the use of robotic seal pups to comfort dementia patients, and a growing number of psychiatric conditions are being addressed therapeutically through automated screen-based characters -- in other words, social robots.
| }
|