There is little doubt that IBM’s Watson artificial intelligence system is an incredible piece of technology, It’s capable of searching across vast repositories of unstructured digital data and returning answers remarkably quickly. But it won’t be replacing humans anytime soon, Dr. John Kelly who leads the Watson team at IBM told TechCrunch’s Alex Wilhelm today at TechCrunch Disrupt in San Francisco.
Instead, the system will augment humans and help us make better decisions, he said.
To show off Watson’s pure intellectual power, the interview began with a demo. Wilhelm chose from a series of questions asking Watson about the accuracy of Wikipedia, the online encyclopedia. In just a few seconds, Watson then reviewed all of the articles on Wikipedia to determine if it was a reliable source or not. After reviewing the body of evidence, it concluded that it was in fact an accurate source of information.
This kind of capability, complete with a computer-generated voice response right out of Star Trek, could be available to developers in 12-18 months, Kelly told Wilhelm.
When people typically think about Watson, they probably remember the time Watson beat the greatest Jeopardy champions of all time, but Watson is much more than that today. Kelly says when they created the system for Jeopardy, it was a pure question and answer engine.
Today, it can ingest all of the information as it did with the Wikipedia demo, or any large body of information. So for example, it could review all of the available literature about a given disease, then work alongside doctors to help come up with the best treatment options. This isn’t about the machine replacing the doctor, but about helping the doctor make better decisions.
Watson At Your Service
When IBM developed Watson for that 2011 Jeopardy! appearance, it was a powerful combination of hardware and software. Today, the company has taken all of that machine-generated intelligence and created a series of services for developers.
Kelly said the company could have built lots of Watson boxes, but they determined putting it in the cloud and making it available via APIs was the best way to distribute the technology.
“Think of Watson as whole series of statistical learning engines, not one thing,” Kelly said. “The interesting thing is it’s so powerful, we don’t have to give a sales pitch.” Developers want to work with these tools.
Just yesterday at the TechCrunch Disrupt Hackathon, teams were using tools in Watson to create hacks, including Kno, an intelligent news finding app
Kelly admits the company is still trying to get past the stodgy ‘Big Blue’ image and attract developers and investors to the Watson platform. While he says, the number of developers is growing exponentially, it has still not completely won over the Silicon Valley community. He hinted that they were working on something to fix that.
The system has vast capabilities. The challenge will be trying to focus it on specific business problems as the company is trying to do with Watson Health and Watson Finance to put solutions to bear on specific industry problems.
Regardless, you can’t question the technical underpinnings of this system.
“This is brand new technology. It’s going to change the world. Think about making decisions without bias or incomplete information,” he said.