Emma Wall: Hello and welcome to Morningstar. I'm Emma Wall and I'm joined today by Jeremy White to talk about Artificial Intelligence. Hello, Jeremy.
Jeremy White: Hello.
Wall: Just a bit of buzz word at the moment AI, but what does it actually mean?
White: It's a good question. It actually means many different things and people use it as a catch-hold to try and workout what they are talking about really, because it includes machine learning and it includes narrow AI, and broader AI.
So, most people have encountered narrow AI what they consider to be with these chat box that you talk on your phone and you think you're having an conversation with somebody but you're not. That's very narrow. It's almost like a decision pre-process you can very easily fool them.
Where we want to get to is conversations with computers or a network of computers that are actually feel like conversations. And one of the easiest ways to explain where we are going with artificial intelligence which is basically just teaching a neural network of computers or a computer to learn on its own is through visual identification or through image processing. We're seeing that all the time.
So, what they – people or companies like Facebook have been doing is to try and teach computers how to recognise objects classify them, identify them and then put them in context and that's the important thing. So, their systems not only get as good as identifying cars and they make the model of the year, but they can also say things like this car is on the driveway of this house.
And if you have ever dealt with a two or three-year old child those are actually quite complex sentences and you got to remember this is the AI determining that for you. And so, when these get more and more sophisticated that's when you start to see company's start using AI for example to identify felons or at football matches which has happened in U.K. or in China where in stating this identifying people with outstanding warrants and things like that.
So, humans cannot actually do that at scale and speed what AI's can. Artificial Intelligence is essentially way of crunching enormous amounts of data down into very, very simple things that you can understand. It does the heavy-lifting so you can then work out what to do with it.
Wall: You've mentioned there a couple of applications which are on the side of kind of civil unrest. What about for the corporate world not just in terms of financial services as an investor which companies are using it in a beneficial way to drive efficiency and therefore share prices?
White: Well actually there is a really good one a company called Orbital Insight which uses combines artificial intelligence and satellite imagery to predict stock performance of companies. So, it monitors the number of cars in the car parks, in shopping malls, and it monitors oil reserves, water reserves, agricultural performance and that's a huge amount of data.
Again, crunching that enormous amount of data and using AI to actually simplify it so you can understand it. And they were successful in predicting a rather large percentage downturn in cars at JC Penney shopping malls in the states and predicted the coming which then actually did take place the coming decline in stock price in the third or fourth quarter last year and then they ended up closing something like a 130-stores. This was something that Orbital Insight used AI to predict before it actually was announced.
Wall: Jeremy thank you very much.
White: Welcome.
Wall: This is Emma wall for Morningstar. Thank you for watching.
