How Real Is “Her?” Five Myths About Intelligent Assistants.
Warner Bros. Pictures
Spike Jonze’s film “Her” — which won the Academy Award for Original Screenplay on Sunday — has sparked a flurry of discourse about the forms intelligent assistants of the future will take. As CEO of Expect Labs, which is building a platform to power a new generation of intelligent assistants, here are the top five assistant-related myths I’ve heard — and the reality of what we can expect in the future.
Myth 1: Intelligent assistants have emotions
If you’ve ever seen “Star Trek,” “2001: A Space Odyssey” or “Her,” it might appear that Hollywood is brainwashing us to believe that the pinnacle of machine intelligence is a computer’s ability to emotionally connect with humans. This idea, while it does serve as a captivating plotline, is not necessarily true. Anyone who has seen someone fall for a “synthetic partner,” or witnessed their child falling in love with a doll, knows that intelligence and emotion aren’t necessarily linked. Inanimate objects are not smarter because they can capture the heart of a flesh-and-blood person. Intelligence is not a two-dimensional experience; your intelligent assistant can still be smart, even if you don’t fall head over heels for it. As long as software can accomplish a given task, it should be deemed effective. We need to put an end to personifying our assistants just because we want another thing to pin our emotions onto. Technology is intended to dissolve the difficulties in our lives — not as a means to acquire adorable little sidekicks. Sure, we can like our assistants, but love should never be part of the equation.
Myth 2: The first place we’ll use them is on our phones
A smart smartphone assistant is not the gold standard. If you’ve ever waited for eons on a customer-support line just to ask a simple question, you’d probably give up all of your limbs for the most unintelligent assistant to come to your rescue. We can do better than this. Intelligent assistants will really come into their own when they’re used in narrow domains. This trend is already happening with the inception of in-car digital assistants: Mercedes-Benz unveiled its Prediction Engine at this year’s CES, and Hyundai recently announced that it is baking an assistant into its new Genesis car. It’s far more practical to have an intelligent assistant that can accurately anticipate and answer your questions about health policies and road conditions than having one on your phone that cannot even muster enough knowledge to do one thing right. Specific realms are where we’ll truly see intelligent assistants become most assistive.
Myth 3: They have to talk back to you
Anybody who has ever worked with an overly gregarious co-worker knows that sometimes the best trait is to be seen and not heard. It can be unnatural for objects to engage in two-way conversations with us. At their most basic level, assistants are meant to assist us. If digital assistants are busy devoting their precious computer brainpower to being good conversationalists, it detracts from what they were set out to accomplish. In “Iron Man,” Tony Stark’s assistant, J.A.R.V.I.S., which stands for “just a rather very intelligent assistant,” is both a helpful aide and best friend. But J.A.R.V.I.S.’s constant babble takes up a lot of its master’s time. If an intelligent assistant doesn’t understand a question, the best scenario would be if it presented users with a list of choices that they could then whittle down in a few simple steps. Conversation does not always equal intelligence. Who wants objects talking back to them, anyway? You have your kids for that.
Myth 4: We will all have our own omniscient personal assistant
While it is the stuff movies are made of, having one piece of software that is responsible for everything under the sun is not sustainable. In the coming years, there won’t be a universal personal assistant that can flit between each job in your life. Instead, each service we use, from an app to a website, will have an expert concierge housed inside. I do think it was really sweet that Samantha was able to submit Theodore Twombly’s manuscript to a publishing house and compose brilliant soundtracks all in the same breath — but that’s not realistic. In the future, we’ll be interacting with hundreds of personal assistants that are specifically optimized for each task that needs to be accomplished. For example, a symphony’s personal assistant that has an expert understanding of musical composition would assist with writing that next masterpiece — not a Samantha. When you go to the hospital, triage won’t be done by a doctor or the individual assistant on your phone, but by the hospital’s own hyperintelligent assistant. Armies of different intelligent assistants that operate with precision and finesse will trump master assistants that can barely do one thing right.
Myth 5: Intelligent assistants will make us stupid
Imagine a future where intelligent assistants siphon us away from the mundane details that cloud our attention and direct us toward the most thought-provoking issues at hand: A cure for cancer, how to design the strongest bridge span, etc. Banal tasks, like making restaurant reservations, calculating tips, or conducting background research, will be a thing of the past. Back in 2007, Wired’s Clive Thompson wrote that “the perfect recall of silicon memory can be an enormous boon to thinking … the machine helps me rediscover things I’d forgotten I knew.” Could using intelligent assistants actually intensify our knowledge? Some assistants even go beyond simple “silicon memory” recall and can predict the information we need before we realize it ourselves. This ability to outsource menial tasks and proactively discover relevant content will free us up to live better. Intelligent assistants will — dare I say it — make us more intelligent.
Timothy Tuttle is the founder and CEO of Expect Labs, a technology company that is building a platform to power a new generation of intelligent assistants. His first company, Bang Networks, built the Internet’s first large-scale CDN for real-time data; his second company, Truveo, built the Web’s second-largest video search platform, reaching more than 70 million monthly visitors, and was acquired by AOL. Reach him @tim_tuttle.