Video: NexChange Talks AI With MIT’s Will Knight
Andrew Work speaks to Will Knight, Senior Editor for AI MIT Technology Review, was the MC for the Emtech Hong Kong Event hosted by MIT Technology Review and Koelnmesse. He covers the latest advances in AI and related fields, including machine learning, automated driving, and robotics. He is also a frequent guest on TV and radio shows.
NEXCHANGE: Hi I’m Andrew Work here for NexChange and I’m at the EmTech event in Hong Kong. I am speaking to Will Knight who is the Senior Editor for Artificial intelligence with the MIT Tech Review. And Will, really excited you can take a little bit of time out to speak to us today.
Will is at the heart of possibly the world’s greatest center for artificial intelligence and robotics but you’ve been traveling to China. Even in Beijing, you’ve been in Shenzhen and you were telling me earlier you’ve seen some things that are really excited you. I imagine you’re a hard guy to impress. What did you see that you thought was so interesting?
WILL KNIGHT: I’ve been getting more and more of a sense that China is starting to produce some really interesting AI companies and so I came here to investigate that. And what I’ve been seeing is that not only are people in China coming up with some great business models around advances in machine learning but especially the big companies are starting to do really fundamental important research. And I think that’s going to be a huge trend actually. I think it’s going to be make a big difference to the the future of AI. I mean it’s already a pretty international phenomenon but the amount of data that’s available in China, the amount of interest there is in commercializing this stuff is just phenomenal. I think that the US companies need to look out.
NEXCHANGE: Okay, can you give us an example of one or two things in particular that you saw, either an application or you can go full dork and talk about maybe some basic research that you saw? Do you want to talk about one of those things being commercialized or one of those things that was basic research?
WILL KNIGHT: I can talk about something that kind of touches on both. A couple of things, so I met with the people from Ant Financial which is the spin-off of AliPay, super impressed by how they are applying machine learning to personal finance and mobile finance. And really using it throughout their company. It’s basically an AI company that does finance so their customer service is like 90 something percent automated. You get in touch with them. You’re talking to a machine, sometimes hands it off, but it’s kind of phenomenal what they’re doing.
In Shenzhen, not far from here, I visited Tencent and I’m super impressed by just the hardcore machine learning work they’re doing and where they’re applying that. And one area I think that Tencent has a great incentive and opportunity is in language, in trying to do like genuinely intelligent personal assistance and that’s something I know they’re kind of looking at, thinking about, and committing to anything. But that’s like a holy grail of a lot of AI research is trying to crank language and if you think about it who has more kind of language data to try and apply some learning to than a company like WeChat.
NEXCHANGE: Yeah absolutely so where you are, it’s famous for being one of the world’s great academic centers and then you get a lot of small spin-off companies. But you’ve been visiting big companies in China and looking at the basic research that they’re doing. Did anything strike you that it’s different or there’s a different dimension to the way they do things? Are the way they think about it? Or do they also have that university connection? Or did you see it when you’re there?
WILL KNIGHT: It’s a little different. I think it’s a little more commercially focused as there’s such great kind of ecosystem for entrepreneurship. There’s a lot of people spinning out companies but a lot of sort of entrepreneurs just starting with the business model and working from there. And so it’s a little different but I think it’s funny to change your spending to become a bit closer to the US model. You know some of these big companies are beginning to bring in professors or have professors work part time which is a real clear sign that they have more interest in the fundamentals. Not just how do we apply deep learning to whatever problem, but how do we actually invent the next deep learning?
NEXCHANGE: Okay sounds great! When can we look forward to reading about some of what you’ve seen in the MIT Tech Review?
WILL KNIGHT: So we’re going to have a special issue on AI in November and my piece is going to be in that. I’m sure everybody’s going to be looking for that piece.