It would be hard to argue that that there is a dearth of artificial intelligence platforms trying to get into the healthcare game. But it isn’t just developers deciding where the industry is heading; it's also up to the healthcare players holding the purse strings.
“Investors get a different view into technology than a lot of folks, and the decisions they make about where to put capital have big implications for what kind of technology, medicines, and products get developed,” Meg Tirrell, a reporter at CNBC and a moderator for this week's AI World Medical Innovation Forum in Boston, said during an investor-focused panel discussion.
At the conference, where key investors in the field came together to talk about where they see the future of AI, most agreed that, in general, implementation of AI is in its very early stages and somewhat burdened by hype.
“I just went to HIMSS' conference and there were about 5,000 vendors, and almost all of them said they do AI, machine learning, patient engagement, and blockchain. But when you talk to them, few of them are using AI; most of them are algorithmic,” Dr. Joe Cunningham, managing director at Sante Ventures, said during the session. “I’m a true believer [that] AI is a generational technology that will change our world, but it's very early on.”
Looking towards the future, many of the panelists said AI could help with speed and efficiency. That doesn’t mean chemists and doctors need to fret over job security just yet, but it could lead to improvements in the way people work.
“I’m not as convinced the workforce is going to get eliminated — I think the workforce is going to get much, much better in all the areas our health system operates in," Amir Nashat, partner at Polaris Partners, said during the session. "The consequence of that is that things are going to start moving a lot faster, and whether the health system is willing to have really great radiologists all the time who aren’t going to make mistakes."
Doctor training, drug discovery, administrative tasks, and diagnostics were some of the key areas where investors saw the technology moving in the future.
“We have to change the way we train doctors," Cunningham said. "A big part of that is clinical decision support. One of the ways AI can be used is clinical decision support at the right time, with the right provider, at the same place."
Meanwhile, Vijay Pande, a partner at Andreessen Horowitz, said he wants to see technology that focuses on a repeatable process that is applicable in more than one area. He said that right now, drug discovery feels too much like “winning the lottery.”
But the process could apply in a number of areas. He gave the example of Freenome, an AI company that uses blood sampling to help detect cancer.
“Freenome built not a test but a process to build tests, such that if they had an ovarian cancer test they can build a breast cancer test by getting more data samples. That is what makes it a repeatable process,” Pande said.
But when it comes to getting the investors to back a company, the panelists were in firm agreement.
“[Investors want AI that] is solving a problem in a unique way,” Lei Yang, managing director at Northern Light Venture Capital said. “Initially I don’t really hope that that the entrepreneur is solving the biggest problem, but a problem.”
Yang went on to say that starting with a smaller issue can help developers get in the game and see where new opportunities are in the future.
In some a crowded field, though, whether investors bite will also depend on who the innovators are who are pitching the product.
“It’s [about] solving an impossible problem but doing it with a really, really savvy group of people, because for us right now AI is a term everyone is stamping on their decks," Nashat said.
