What is the Internet of Things?
When we talk about the Internet of Things, what we mean is computers gathering information by themselves. In the 20th century, in the 1990s, computers only got data from people, typically people using keyboards. In the 21st century, it has become possible for computers, and therefore machines, to understand the world around them by themselves. Your smartphone has about 10 sensors – it knows where it is, it knows which direction it’s going, it knows whether or not it’s moving, it knows what the ambient temperature is, it knows what the barometric pressure is. You don’t have to tell it any of these things. And that allows a number of applications.

One Internet of Things application, which a lot of people don’t even think of as an Internet of Things application, is Uber, the ride-sharing service. Uber could not work if people did not have smartphones that knew their location. You can see why the internet is important in the Uber application – because your smartphone knows where you are, there’s somebody willing to give you a ride, their smartphone knows where they are, that information goes into the internet, and at Uber there’s a piece of software that finds the nearest driver for you and gets you there. That’s a very simple example of how internet-connected sensing delivers value.

How was the term born?
‘The Internet of Things’ was just the title of a PowerPoint presentation I made in 1999. I was a junior executive at Procter & Gamble, a huge multinational consumer goods company from the US, and I had figured out that to solve some of our supply chain problems, we really needed to put little microchips in all our products. And I had to explain that to senior executives who didn’t have very much time, who weren’t very technical, but I knew that they understood that the internet was a big deal in the 1990s and they were very excited about it, so I wanted to get the word ‘internet’ into the title of my presentation to grab their attention. That’s how ‘The Internet of Things’ came about. At the time I don’t think anybody was talking about the internet of anything, it was just ‘the internet’. I probably should have said, The Internet for Things, which would have been more grammatically correct, but I said The Internet of Things. The presentation was successful and they gave me some money to go and do some research at MIT, and I gave the presentation again and again for many years, and I guess the term just stuck. So it was a very accidental thing.

What has changed since 1999?
The Internet of Things has unfolded in some ways much faster than I thought it would, and in other ways much more slowly. We got a lot of things right in the late 1990s and early 2000s about the importance of having computers able to gather their own information and the value of sensing things automatically. One thing that people use every day – and no longer even think about – is the GPS. It’s almost hard to remember how we used to get around without having a kind of map in our car that was interactive, that knew where we were. We’ve taken that very much for granted, but that’s an example of computers sensing something for themselves – in this case, location – and then delivering a very useful application with that information.

We got things right where we saw that it would be very valuable to have automated sensing. What was completely unforeseeable in the late 1990s and early 2000s was the rapid advance of wireless technology. There was no WiFi – to get on the internet, you had to plug in. Cellular telephony was really quite basic and wasn’t very good for data at that time, so the incredible advances in those kinds of technologies were far faster than we expected.

In other areas things have been a lot slower. I predicted many times that we would have these radio frequency identification tags in consumer products by now. They do exist in some products, are used in some applications, but the consumer goods industry has been very slow to adopt this technology, and it’s not for technical or economic reasons – it’s really for political reasons. It’s figuring out who pays and who gets the data, stuff like that, that’s slowing things down. I think the overall lesson is that conceptually, we got a lot right back in the 1990s and, if anything, the technology improved much faster than we thought it would, but some of the human factors have taken a lot longer to unfold than we expected.

Where will the biggest disruptions happen?
The Internet of Things is really a 100-year project, and we’re kind of into the 17th year. I don’t think it’s going to be a linear progression, but certainly the impact that the Internet of Things has will continue to grow. What’s coming next – the big, visible transformation everyone will notice – will be in transportation. We’re going to see a rapid deployment of self-driving vehicles, both passenger vehicles and commercial vehicles like trucks, in the next 20 years. For me, the interesting thing about this development is what happened in the past – when you look at history, if you see changes in transportation, you see massive changes in society and in geography, in how human beings use land. So we have a really exciting couple of decades coming.

How will we manage all the data?
When I think about the Internet of Things, I don’t think of devices so much – I think of sensors. Every smartphone represents eight to ten sensors, and there are about 10 billion smartphones in the world, so that’s 100 billion sensors right there. That means a lot of data. So right now, data analysis is really the most interesting aspect of the Internet of Things. We need more and more software that will analyse that data for us, using a technology that some people call artificial intelligence, although I prefer to call it machine learning. This means having a system look at data, find patterns, figure out whether the patterns it has found are meaningful, and if they are, learn to do more of that, and if they have false positives in them, learn not to make that mistake again – that’s what we mean by machine learning.

The only way to handle this data coming from the Internet of Things, and all these tens of billions of sensors, is creating better and better algorithms, better and better software. But that’s one of the hardest things right now because there are so few people who know how to do it. People who are experts in machine learning are in huge demand. As more machine-learning experts enter the workforce, and as we learn more about how to analyse data algorithmically, we’ll see better and better ways of turning big data into big value.

What about our privacy?
Whenever you talk about the Internet of Things, privacy is one of the first things people want to know about. It’s worth understanding what I mean by privacy. Privacy is the ability to determine who sees your information. You may want to share some information with some people and not with the world, or not with certain people – that’s what privacy is, it’s discretion around data. We have a couple of ways to provide people with privacy. The first one is helping them to understand what data is being captured and how it’s going to be used. If you don’t do that, you’re not giving people the choice of how their data is used, which is a big problem right now. Take those home microphone systems, for example, where you ask a device to order something for you, say milk, and people don’t realise that those microphones are listening all the time. They have to be listening all the time, because otherwise they won’t hear you when you tell them you need some milk.

So, communicating what a technology does, and what happens to the information that it gathers, is very important. The way this is handled today is with these complicated terms and conditions – whenever you want to use any kind of data service, you have to check a box that says you have read the terms and conditions. You didn’t read the terms and conditions – nobody reads the terms and conditions! – but if you did, you wouldn’t really understand what they are trying to tell you anyway. They’re designed to be obscure.

So the first thing we need is a kind of privacy label on a product, a little bit like the nutrition label on a food product that tells you about the calories and ingredients. We need a very simple, easy-to-understand and clearly regulated way for consumers to understand what data is being gathered by which technology they use.

Beyond that, there’s another problem – even if I promise to keep your data private, I have to be able to deliver on that promise with good security technology that is properly implemented. But that’s a huge challenge for a lot of governments and corporations today. The technology exists to keep data secure – the technology exists to keep data very, very secure – but a lot of corporations and governments don’t use it. They don’t implement it properly, they take shortcuts, and those shortcuts lead to vulnerabilities.

So to my mind, everything around privacy and security is in the realm of policy – either government policy or corporate policy. It’s a matter of ‘Are you doing the things you could be doing to keep people’s data private?’ Sooner or later the organisations that don’t do it are going to suffer because people won’t buy their products or won’t vote for those governments any more.

The interview with Kevin Ashton was conducted in last October during the yearly Generation Y Conference in Barcelona, which was dedicated to the topic of ‘Big Data’.