“Data is the new oil” may be a classic cliche characterizing how important raw numbers are for the computer industry, but when it comes to artificial intelligence ventures, the cliche may not go far enough.
“One of the big blocks for AI is data,” Ben Wilson, director of the Center for Intelligent Devices at Bellevue, Wash.-based Intellectual Ventures, said today at a forum about AI presented as part of the Seattle Metropolitan Chamber of Commerce’s Executive Speaker Series. “Traditionally, startup companies need capital. Now, if you’re doing AI, you need capital and you also need data. And you’re going to burn through your data before you burn through your capital.”
Wilson pointed out that the big players in the AI market are the companies that have the data, whether it’s Amazon or Microsoft, Facebook or Google.
“Before you have a good idea, start with data,” he said. “And if you’re someone who has a great idea but you have no data, that’s going to be a big roadblock for you, and you’re going to have to find some collaborators or partners who have access to the data you need.”
Data sets are of paramount importance because the tools for machine learning and computer vision have a voracious appetite for training data. Gigabytes upon gigabytes of examples are needed to teach an AI agent to distinguish between a cat or a dog, or between a stop sign and a pedestrian walking across the street.
“Specifically with machine learning and deep learning, data is the new currency that you have to consider when you are doing business plans,” Wilson said.
Sridhar Chandrashekar – co-founder and CEO of Optio3, a startup focusing on enterprise management software for the Internet of Things. or IoT – seconded that view. He noted that in his talks with companies such as GE and Honeywell, customers are often interested in just buying the raw data for data’s sake.
“So there’s your business opportunity,” he told the luncheon crowd. “Go into the business of collecting data and making it available to AI companies.”
This year’s acquisition of Seattle-based Mighty AI by Uber serves as an illustration that’s ripped from the headlines. Mighty AI’s business model was to develop training data for the computer vision models used in self-driving cars, and Uber clearly decided it was cheaper to buy the company than to purchase the data piecemeal. (Terms of the acquisition were not disclosed, but according to one GeekWire source, investors didn’t get back all of the $27 million that had been raised for Mighty Ai.)
Seattle stands out as an AI hotspot, largely due to a talent pool of engineers cultivated by the likes of Microsoft and Amazon as well as research institutions such as the University of Washington and the Allen Institute for Artificial Intelligence. And we’re not just talking about computer science and engineering majors. Chandrashekar said mathematicians and physicists are in the mix as well.
“One of the best engineers you can find for machine learning is from the field of statistics,” he said. “They actually think about data like nobody’s business.”
Thurman said his university offers more than a dozen joint degree programs that blend computer science and other disciplines – ranging from math and science to the humanities. “I can tell you what employers are telling us they’re looking for,” he said. “It’s all of that stuff, and people who have the ability to carry on a conversation with each other.”
Here are some other highlights from today’s conversation, moderated by Seattle Times reporter Melissa Hellmann:
- Thurman said AI isn’t just for tech titans anymore. In one sense or another, every company is becoming a tech company. “We place as many students at Nordstrom as we do at Microsoft,” he said.
- The panelists said worker retraining will be key to addressing the shifts in employment patterns that will crop up due to the rise of AI and automation in the workplace. Wilson pointed to the Lifelong Learning and Training Account Act, a measure that would set up government-matched savings accounts to support workers as they learn new skills. Rep. Suzan DelBene, D-Wash., is one of the bill’s co-sponsors.
- Wilson worried that AI’s rise could sharpen income inequities unless they were balanced by public policy initiatives. “If you look at who is best positioned to capitalize on AI … people with data are in a position to gain a lot of money, and those tend to be the people who already have a lot of money,” he said.
- How about requiring companies to pay a person for the data they collect? “One way a lot of people have proposed to solve the income inequality problem is for people like us to realize that the data we generate is valuable, and make people pay for it,” Wilson said. “The hard part is, you give your data away for free everyday to everyone. … It was one of the last revolutions of the internet, and maybe we have to undo [that] to solve some of these problems.”