IBM Watson's Goodbye To Privacy
IBM held an event in San Francisco recently to show off new capabilities in Watson, it’s artificial intelligence system that’s being made available to developers to let them build smarter, "cognitive" applications.
To set the futuristic tone, IBM invited Peter Diamandis, founder of the nonprofit X Prize Foundation, which humbly describes itself as "a catalyst for the benefit of humanity." To give you an idea of Diamandis' interests, he said he is currently "prospecting" asteroids that he plans to mine for resources. He put the value of one asteroid at $5.4 trillion.
But he was there to talk about Watson, and how humankind is producing so much data these days that we can no longer make sense of it all without artificial intelligence (AI). Watson and its ilk are needed to uncover patterns in mountains of information and make decisions we can no longer arrive at through traditional programming.
This isn’t big data it’s gargantuan data. Take, for example, images from new fleets of satellites that can see things as small as 50 centimeters across.
“You want to know what your competitors in China are doing? You can watch them,” said Diamandis. Want to predict what Best Buy’s earnings will be at the end of the quarter? Count the number of cars in its parking lots and the size of packages people are carrying out. "Well, you couldn't count them," he said, "but Watson could.”
It’s a fantasy, and yet not really. That data is starting to be collected, and IBM, along with rivals like Google and Microsoft, are building AI systems that will make that type of analysis possible. But the data gets closer to home than satellite images, and when it does there's little room for privacy.
“We’re heading towards a world of near perfect data,” Diamandis said, where “perfect” means everything that happens is recorded and available for someone to mine.
Self-driving cars, he notes, are constantly monitoring their surroundings, capturing image data at 250MB per second. “You’ll never have an accident and not know what happened,” he says.
Add to that an army of miniature image-snapping drones cruising the streets, and "you’ll never get pick-pocketed and not know who did it."
“If you thought your privacy wasn't dead yet,” he concluded, “think again.”
If that doesn't sound thrilling, there's a trade-off for giving up all this data, a whole new world of convenience. And it’s already starting to emerge in the form of 100 applications that developers have built using Watson’s cognitive APIs.
One of them on show at the Watson event was a hotel concierge app like you've never seen.
The Ivy app from Go Moment uses Watson's natural language APIs to interpret the messages you type, and it's been loaded with everything there is to know about the hotel you're staying in and the surrounding area, including restaurants, theaters and other services. There are no more calls to the front desk and waiting an eternity for someone to show up at your room. Need tickets for a show and a beer and some towels by the pool and recommendations for dinner? Ask Ivy and it shall deliver. Watson's APIs basically allow any app developer to create their own version of Siri, with knowledge about any environment they load in.
Ivy will be available to 20 million hotel guests by the end of the year, according to Go Moment CEO Raj Singh. Hotels love it, he said, because it keeps guests happy and cuts labor costs. That's because it can field most questions itself, and those it can't it routes directly to the right department without tying up the front desk. "We automate two months of labor every day." he claimed.
Another Watson-powered service is WayBlazer, which is basically Expedia on steroids. The service lets you input searches using natural language -- "What's the best hotel for a relaxing extended weekend?" -- and spits out results based on a profile it builds over time. The data is culled from thousands of sources including social media, blogs, magazines and newspapers.
CEO Felix Laboy arranged a "girls weekend" for one of IBM's executives, which included spas, massages and plenty of shopping. It seemed out of touch with Silicon Valley's current obsession with diversity and equality, and the executive pointed out that she might have enjoyed some outdoor activities as well.
Marketers are another big target for Watson, and former Sun CEO Scott McNealy was there to show his service, WayIn. It uses Watson's image recognition capabilities to trawl photos on social media and make them searchable, even when they don't have tags describing their content.
"We'll ingest and tag 200 million pictures a day," he said, which can be filtered by demographic and other attributes. "You can't do this without Watson."
