by Peter High, published on Forbes
Over the past decade and a half, Microsoft co-founder, Paul Allen, has created three “Allen Institutes” for Brian Science, Cell Science, and Artificial Intelligence. The Institute for AI was founded in 2013, andits mission is “to contribute to humanity through high-impact AI research and engineering.”
In early 2014, Allen tapped serial entrepreneur, Oren Etzioni, as chief executive officer. Etzioni has a PhD in computer science, has been a professor at the University of Washington, and founded or co-founded a number of companies, including Farecast (sold to Microsoft in 2008) and Decide (sold to eBay in 2013).
The goal of Etzioni’s research is to solve fundamental problems in AI, particularly the automatic learning of knowledge from text. In our far ranging conversation, we discuss the specifics of his goal, the pace of innovation in AI more generally, safety concerns, and how they should be dealt with, the government’s role in mitigating risks of AI, and a variety of other topics.
(To listen to an unabridged audio version of this interview, please click this link. This is the fifth interview in my artificial intelligence series. Please visit these links to interviews with Mike Rhodin of IBM Watson, Sebastian Thrun of Udacity, Scott Phoenix of Vicarious, Antoine Blondeau of Sentient Technologies, and Greg Brockman of OpenAI.)
Peter High: You are the CEO of the Allen Institute for Artificial Intelligence whose mission is to contribute to humanity through high impact AI research and engineering. Can you provide your definition for high impact AI research and engineering?
Oren Etzioni: It starts with Paul Allen, who is a visionary and scientific philanthropist. He won the Carnegie Medal for Philanthropy last year. He has been passionate for decades about AI research and the potential of AI to benefit humanity.
In January 2014, we were launched as a nonprofit research institute in Seattle. We are now fifty people – about half PhDs and half engineers – and the question that we ask ourselves when we get up in the morning is “What can we do using the techniques?” Ultimately, to me, the computer is just a big pencil. What can we sketch using this pencil that makes a positive difference to society, and advances the state of the art, hopefully in an out-sized way? We are small compared to the teams that Google and Facebook and others have, but we want to punch above our weight class.
One of the things we have noticed as we have developed expertise in natural language processing and machine learning is that there are millions of scientific papers published every year – nobody can keep up. Google Scholar came on the scene about a decade ago and indexed all these papers, but there is too much information: You do a simple query and experience overload. What we need are techniques to help people cut through the clutter and hone in on key results. The approach we have taken is to use AI methods to filter irrelevant results—to extract key information like the topic of the paper, the figures that are involved, the citations that are influential, etc., etc.— in order to help people find the papers that they need. We have launched a free service on the internet called SemanticScholar.org, which currently indexes several million computer science papers. Our hypothesis is that if we can make scientists better at their job, then we can help solve some of humanity’s thorniest problems. We are starting with computer scientists, but we want to expand to medical researchers and ultimately doctors. Even a specialist does not have the latest information about your condition– they just cannot keep up. They are diagnosing you and treating you based on, at best, incomplete and potentially erroneous information. We want to help change that.
High: If you were to think about the next decade, what are some of the promising future attributes outcomes that you foresee with the developments that are coming down the pipeline and with regard to AI generally speaking?
Etzioni: AI is becoming pervasive in its use in technology in society. Marc Andreessen famously said that software is eating the world. One might riff on that and say that AI is eating software, in the sense that everywhere where there is a software solution, there is the potential for an AI solution.
Cars are a great example: They have become complex computers. There are more than one hundred fifty computers in the average car. There is the potential now to have a car drive itself using AI. The reason that is exciting is that it could reduce the number of accidents we have on the roads today due to distracted human drivers or humans driving under the influence. Our highways and our roads are underutilized because of the allowances we have to make for human drivers. We could pack the roads a lot more densely and reduce traffic congestion and greenhouse gases and all those things if traffic were more efficient, so that is a great example. But, anywhere you look in society I see the potential for AI to help.
High: I read a paper of yours from a number of months back in which you said, “The popular dystopian vision of AI is wrong for one simple reason: it equates intelligence with autonomy.” I wonder if you could unpack that insight a little bit.