📚 Check out our new AI & Data Reading List: Dive Deeper! ✨

AI Today Podcast #003 – Guest Expert: Oliver Christie: “Have predictions for AI in 2017 Come True?”

Podcast #3_ Guest Expert_ Oliver Christie “Have predictions for AI in 2017 Come True_”

Podcast #003 – Guest Expert: Oliver Christie: “Have predictions for AI in 2017 Come True?”

Show Notes:

On today’s show we interview guest expert Oliver Christie and discuss his 5 predictions for artificial intelligence in 2017.  Now that we’re about two thirds of the way through 2017 we discuss what happened with his predictions – which ones came true, which ones are still in the far future.  We also discuss the relationship between big data and artificial intelligence.

  • Article in Medium: 5 Artificial Intelligence and Data Predictions for 2017
  • Prediction #1. Insight Data will Replace Big Data
  • Prediction #2. Data in a Post-Truth World
  • Prediction #3 . The Public Demand For Information.
  • Prediction #4 . Data Will Become Connected
  • Prediction #5 . Companies Not Using AI Will Get Left Behind


A transcript of podcast is available below:

Kathleen Walch: [00:00:04] and welcome to the AI today podcast. I’m your host Kathleen Walch.

Ron Schmelzer: [00:00:09] And I’m your host Ronald Schmelzer. Our guest today is Oliver Christie and a consultant and expert in the field of artificial intelligence. Hello Oliver.

Oliver Christie: [00:00:19] Hello, good morning. Very nice to meet you both.

Kathleen Walch: [00:00:22] Likewise. We’d like to start today by you introducing yourself to our listeners and telling us a little bit about the things that you’re doing in the field of AI.

Oliver Christie: [00:00:34] Yes absolutely. So I’ve been in the field of AI some time. Mostly centered around the natural language processing and human insights. What’s the crossover between people and how we act and how AI can understand that better. I’m a consultant so I’d say probably half of my work is done with tech companies and big financial institutions so they can figure out how to implement AI. Then the other half the work I do is research based. Think about where AI is going what it’s going to look like next. And then also how a company might look if it’s re-built from scratch with AI and data. At. The call.

Ron Schmelzer: [00:01:30] Great, well obviously we’re currently in the midst of all this transition to artificial intelligence so even though this has been something been going on for the past few decades with development around artificial intelligence obviously we’re still at the very beginning. So on that note you recently wrote a piece in Medium called Five AI and data science predictions for 2017. So I think for our listeners we’d love to have you highlight a few of those and tell us where you see the field of AI going this year. So if you want to kick off the first prediction you made or do you want to work through that yourself.

Oliver Christie: [00:02:08] Sure. The first prediction I made was that big data which was quite a loose term anyway would be replaced with insight data. And what I meant by that It’s it’s not particularly useful having an awful lot of data. unless you know what that data’s about. I can have any quantity of data in front of me but if I don’t know the function of that and how it is useful the data doesn’t really have much purpose. I think the insights are going to be the things that people really start to talk about a lot more. That part is the valuable part anyway.

Ron Schmelzer: [00:02:51] That’s good. Yeah I know we’ve always heard about that hierarchy of data knowledge and information and the differences between them. So you know obviously we’re moving from pools and piles of big data to more information and hopefully from that we’ll get some knowledge. So that definitely sounds key. So in 2017 I know we’re sort of already two thirds of the way here. I mean have you started to see that sort of transition to people using AI and more knowledge base systems to gain more information from their data?

Oliver Christie: [00:03:23] It is happening. It’s not happening as quickly as I think. Not as evenly as I think it was going to. I would imagine all companies with great data will say well what’s the value of what was got here. And we really found out what the underlying message of data. Instead it seems some companies are rushing ahead and doing a fantastic job of it. But an awful lot aren’t. An awful lot are still not quite getting to what’s the part that matters here. Totally different across different industries. So something like financial or advertising are way ahead of industries like Beetle which is somebody behind governments as well. So I feel industries still need to catch up. But it’s getting there.

Kathleen Walch: [00:04:25] OK. Now would you like to talk about another prediction. That you made for 2017.

Oliver Christie: [00:04:34] Yeah so prediction #2 was that how we think about data and post truth world is is changed. If we, if we, now have the idea that scientific fact or an industry view can can be challenge and be challenged by. A large audience. Easily. Well it doesn’t change the fact that that scientific fact is, on the whole, accurate and right and correct and so on. But it’s not there’s a disconnect between the scientific community and the general public at large. And I think that gap has to be wide, has to be narrowed, it’s very wide at the moment and it’s growing. And I think one way of changing that would be to try and review start and try. And. Be in the same arenas of the general public are interested in. Something like climate change is an obvious. It’s obvious to see the disparity. But. If we look at really any type of scientific or technical or industrial or industry view it’s not necessarily connecting and I think the wider community needs to do a better job at this.

Kathleen Walch: [00:06:03] OK, and now you know like Ron had said earlier since we’re about two thirds of the way.through 2017 how have you seen that prediction come true or not so far.

Oliver Christie: [00:06:18] I love to say that we made a change we’ve made it that there’s there’s progress. I don’t think we’re in a particularly different position than we were six months ago or even a year ago with the sort of rhetoric hasn’t changed particularly. The seesawing of the scientific and data communities are not really being heard in the mainstream media. It feels like we’re still in our own bubbles. We’re still talking to our peers but not the wider audience. I don’t think there has been much progress.

Kathleen Walch: [00:07:02] Now, can you explain to me what you mean by post truth world.

Oliver Christie: [00:07:09] In that the things that we assumed were true, the some of the grounded facts that we were taught growing up on the whole quite a few of those are being challenged.

Kathleen Walch: [00:07:28] OK so you’re not gonna fake news here you’re talking about about things that people thought were the truth, but data has proven otherwise.

Oliver Christie: [00:07:39] Well it’s a little bit of both. Fake news certainly falls into that category of defined truth. I mean there are things that data can challenge which is fantastic but I think it’s mostly the presentation of fact. And alternative facts, well you cant have an alternative fact. It doesn’t make any sense. And that’s where I think we are at the moment.

Ron Schmelzer: [00:08:12] Sounds good. Yeah definitely,I know that when you have systems that are trying to you know be built on the idea of knowledge and understanding in artificial intelligence you have to have a single reference to truth. There has to be an understanding of truth. That’s a really interesting overlap between what’s happening in the field of AI what’s happening sort of in this general public environment. And moving on a little bit related to that one of your third prediction is that the public demand for information will increase. So would love to kind of get your feedback of your prediction about that for 2017 and in kind of where we are with that now.

Oliver Christie: [00:08:51] Yeah it’s something I’ve been saying for some time is as humans we’re hungry for information, information on what to buy, what to do, where to go especially for something medical. So what treatment should I get. And you’re hungry for the knowledge so you can make the best choice. Now. The internet is fantastic for providing information but it’s very poor for saying which information is empirically correct. So, anyone can publish online at the moment. And it’s very hard to figure out which things are true which things are not true unless you’ve got a very deep understanding of that particular subject. And even if you have a deep understanding of the subject it’s still quite hard which things can be double blind tested. The public is still looking for information. I think for the people the percentage of the population who are interested in knowledge and understanding this is going to ramp up and the demand is going to be higher simply because with more information we should be about to make a better choice. On the flip side though I think there’s also a fairly large percentage of people who are not interested in knowledge who are not interested in shall, we say understanding, wider facts. And those people tend to get more and more entrenched in that viewpoint. And it’s not necessarily a very right or very left view point. But it’s if it’s a fixed viewpoint and that seems to be harder to change.

Kathleen Walch: [00:10:46] OK. Now prediction number four is that data will become connected. Would you like to elaborate on that.

Oliver Christie: [00:10:54] Yes absolutely. We’ve still got data in silos. We’ve still got individual companies hoarding their own data. A commercial company is medical and so on. But that data in it’s particular silo isn’t necessarily that useful. It will lead you towards one answer for one question. But doesn’t necessarily put things into context. So we’re still stuck with a very narrow. We have a very narrow dataset. With AI we can build off that. It’s still incredibly narrow. We need more data from more places to get a wide understanding of any situation and it’s at that point that AI can really start doing the things that it should be able to be doing. So.

Kathleen Walch: [00:11:55] Yeah, I agree with that. I do think that one of the challenges is you know, data is very proprietary and people hold that you know they protect that. So we have to think of ways that make it beneficial for companies to want to release and share that data with either their competitors or you know across different industries. So that’s going to be a challenge that I see.

Ron Schmelzer: [00:12:22] And you know piggybacking on that you know we’ve been solving this challenge of enterprise integration for decades. I know in my previous analyst business that was actually one of the core focuses focusing on things like service oriented architecture and web services. And of course the movement towards broadly accessible API. That was all trying to solve the problem of the silos of information and people trying to connect to them and integrate the data and just in general trying to make better use of systems. And you’d think that that would have solved that problem. I think we’ve solved that somewhat but it sounds like this is not just the problem of accessing information but trying to access the meta data and the knowledge that they’ve built around that information as well.

Oliver Christie: [00:13:08] I mean we still I still talk to very large banks who say well this is our this is our Web site team here, and this is our team over here who do the app. And then there’s a team member here to do the marketing. And it is confusing as to why we have this approach I mean it’s a technology stack sure but there’s so much overlap between what each team is doing. They should be in the same room. They should be talking about the same thing. And that’s for something as simple as an app. The insight you can get from each team is fantastic. But if they’re not, if it’s staying in that silo things are still siloed, we will lose that usefulness of collected data.

Ron Schmelzer: [00:14:03] And I think that sort of brings us to this last insight that you had which is that companies are not using AI will be left behind and certainly AI is moving at a fast pace. So tell us a little bit more about what you’re actually seeing about companies pulling ahead and getting left behind and just in general about that prediction for this year.

Oliver Christie: [00:14:21] Yeah absolutely. So what I’m saying is that companies who are working on AI now are starting to think about the challenges and thinking about data in new ways. They’re thinking about what the impact is going to be on the structure of their company on the bottom line. All these things. They are getting to the problems quicker. They’re also making progress. But it’s it’s thinking about these big questions. That’s the challenge. But it’s also if you can tackle them. You can solve these problems early you’re way ahead of the competition. And I think the competition won’t be able to catch up. I think once you’ve figured out how AI works in your organization and leverage it to do more than just automation I’m be your competition won’t be able to reach to where you want. So yea I really think we’re seeing somewhat of an arms race.

Ron Schmelzer: [00:15:32] It’s good to know. Just a quick follow up to that. So so what sort of ways do you see them falling behind? Is it just losing market share? Is it losing, you know, competitiveness? I mean so what are the ways that you are seeing AI providing an edge to the companies who are taking advantage of it right now.

Oliver Christie: [00:15:52] We’re slightly too early to say exactly what those advantages are. Quite often it’s more in terms of insight than anything else. There will be a huge financial payoff but that not. It’s there for some companies but not all. But I think instead you’ve got to think of AI not as a simple tool or an add-on or one extra feature. But instead how will a fully connected fully AI driven and data rich company what would that look like. How would it function. And what the real core advantage is that you get from the new technology. They need to think bigger and the companies that are they’re going to be the next Apple and Google and so on. Everyone else is just going to be the next BLockbuster.

Kathleen Walch: [00:16:53] I don’t think they want to be that.

Kathleen Walch: [00:16:56] All right. So, so wrapping it up here we have one last question for you. Obviously there’s a big difference. There’s a big relationship between big data and AI but these areas differ in many respects. So how dependent do you believe the future of AI is on the future of big data and what stumbling blocks in big data will pose challenges for AI adoption.

Oliver Christie: [00:17:18] Yeah so at the moment AI is absolutely dependent on big data. At the moment we still need quite a lot of data to train whatever system we’re going to use to be able to give us good results. So at the moment you might need one hundred examples, I’m sorry, a hundred thousand examples or a million examples to get to a good training model. As we do more training as everyone does more training, we have a better basis for, for where we sit. From there we’ll need less and less data to get good results. Will be on one hand to be much less reliant on big data. On the other hand we’ll start to look at more things inside the data and also think about what we neew data we might want to collect to improve the AI we’ve got.

Kathleen Walch: [00:18:22] So you’re saying that big data is important now. But in the future, it’s, we’re going to be less reliant on it because the systems will have already learned from what we have now?

Oliver Christie: [00:18:33] Yes. So the big data we’ve got is going to be useful for a number of years. And then it’s going to drop off. And we’ll then start collecting a new type of big data. Most likely IoT type device data and more likely sentiment and environmental type data which then this is something richer look at the world.

Ron Schmelzer: [00:19:00] Yeah I definitely definitely think so. I mean obviously if you look at the investments that the major platform companies are making in Facebook, Microsoft, Google, Amazon you know they’re they’re definitely counting on that growing base of data. You know just in general from all the things that they’re encountering from their own customer experiences and their partners and their developers who are building on top of those platforms so that data explosion is has resulted obviously in an accelerated AI landscape and so you know obviously you would think that companies themselves with lots of data would see that same explosion so are were keeping an eye on that. Keeping an eye on what’s happening in the public and private implementation and adoption of AI. Do you have any any final words or thoughts on kind of where we are with the Today and where things are headed in the near future.

Oliver Christie: [00:19:54] Oh think it’s an incredibly interesting and exciting time to be working in AI. I’m based in New York and the amount of things happening here is just fantastic. I’d dare to say 5 years ago or even two years ago. There is a real explosion of startups especially if you’ve got larger companies thinking about it or trying to work out how to use AI and will impact the bottom line and some really interesting questions being asked about where we want to go as a society as consumers and as companies. So it’s a fantastic space to be in right now.

Kathleen Walch: [00:20:41] Okay great. All right well thank you so much Oliver for your time. And thank you listeners. And we’ll catch you at the next podcast. By everyone.

Ron Schmelzer: [00:20:49] Thank you very much for everybody for participating on the AI Today podcast.


Login Or Register


Register to View Event


Get The AI Today Podcast #003 – Guest Expert: Oliver Christie: “Have predictions for AI in 2017 Come True?”


AI Best Practices

Get the Step By Step Checklist for AI Projects


Login to register for events. Don’t have an account? Just register for an event and an account will be created for you!