PODCAST: The Future of AI


0:00 - 0:30

Jason Anderson. I am a chief news editor at The Wall Street Journal ( via WSJ ). I’m speaking with two. Experts in artificial intelligence Orrin at coney is CEO. Oh of the Allen Institute for artificial intelligence its mission is to conduct high impact. Ai Research for the common good that includes projects like a semantic scholar and they I generated public search engine that indexed one hundred and seventy million scientific results from a wide range of disciplines to make access to relevant research easier they also developed Grover a model that

0:30 - 1:00

can generate realistic looking fake news articles in order to be able to detect those fake news articles. Written by in the future Carol Riley is a robot. BADDEST and co-founder of the self-driving car company Dr. Ai. Partnered with the city of Arlington Texas to operate autonomous shuttles. In the area, it was acquired by Apple in June. The big question we want to answer. What is the force that will continue to fuel innovation and artificial intelligence and what systems are guardrails are necessary to protect security and personal

1:00 - 1:30

privacy along the way? Welcome, Carol I thought it’d be good to start since you are both technologists who literally work in artificial intelligence. Just ask for your take on what is as you see it and just as much. What isn’t it and maybe we’ll start? You sure so I would. I’d say I guess three things. The first one is that the origin of the concept back in the fifties was, of course, to emulate and maybe sometimes even surpass human an intelligence. That’s

1:30 - 2:00

not what we’re seeing with the recent surge in interest in machine learning deep learning. What we’re seeing are techniques that utilize realized a massive amount of data to build predictive models that have shown themselves to be effective in a very wide variety of applications for manufacturing to healthcare care and more and more so? That’s the reality of today. And then in terms of what people are thinking about what people have been talking about where we think about the future. I think it’s become almost

2:00 - 2:30

kind of Rohrschack test. People project their own concerns whether it’s about privacy or about taking over or about bias and discrimination jobs. These are all legitimate concerns but I bet you if we did a quiz the professor says okay everybody takes out a piece of paper and writes down four sentences about what you think is and we’ll be and then we’ll compare those there would be quite different. That’s why say Rohrschack care already say I think artificial intelligence

2:30 - 3:00

is a machine displaying some level of intelligence and there’s narrow ai which does very small specific tasks or generally generally I which is the the fear that killer robots coming in and I’ve always worked on the spectrum and I’ve always been very fascinated with the area of human-machine collaboration and I do think that we’re very far off from back narrative of Terminator. But I do a lot of rules and Ways that humans and machines can

3:00 - 3:30

work together and very very broadly. It’s you know I think just to be clear it’s like this big umbrella of Ai and then you have machine learning underneath breath it and then you have deep learning as a subset. So those words are generally interchangeable. But you know it’s kind of like that Larry and I’m excited to see new techniques like deeper learning things that eighteen and even though we’re seeing the term may be a bit overused right now especially from a marketing standpoint. You’re optimistic you’re excited about the technology and you feel like in general the things happening or are truly good and

3:30 - 4:00

exciting. And Oh yeah I think we were. Just now at the cusp of a new breakthrough. So I will say that you know we’ve undergone what I call it the first wave of this transformation. There’s the slogan. A is the new electricity power everything. We’re just kind of started in. Eventually, it will fall into the background. Be So seamless right now. It’s so frightening center in everyone’s minds the first transformation was really powered by data and compute. And you always hear data incomplete. You know and I think right now. We’re entering into this second

4:00 - 4:30

wave of AI. This will be powered by the rise of tools and talent and ideas? Hi, Dea, the really interesting thing is when we talk about like where we’ll be competitive on a global scale in the US or China you know really. It’s this I don’t like pitting against each other and I don’t think it’s a winner. Take all the situation. The popular machine learning course taught by Andrew has taught over a two-point. Five million people

4:30 - 5:00

in that class alone sure and that would have been impossible for any professor to teach no matter how popular you are so. I’m really excited that these tools and these platforms forms are available so that talent can spring from anywhere silicon valley where I’m from used to be like everyone’s like what’s the secret sauce and can we replicate it and we’re a scene like different communities create their own secret sauce because I think so. Many problems are local and Calicut spring from anywhere. Now that you have all these tools available online nine the discipline of

5:00 - 5:30

new group of learners who are learning online the need for ethics. Yeah, so I want to come back. Pack a lot of that. But I’m going to start with one of the things you’ve touched on that. We talked about a lot today. which is this idea of? Ai Is something of an arms race right or wrong. It’s certainly one of the ways. It’s been positioned certainly between countries and I’m going to pull the graphic that is based on some analysis that or are in your group. Did I believe and you can see it here and this is a look at all of the research that was published and the top ten percent of papers that are cited? So I  gather those are ones that tend to have the most impact and

5:30 - 5:36

the ones that are the most prominent and what what you see is a really remarkable increase