The next stage of digital transformation of governments is AI: William Eggers

William D. Eggers, executive director of the Center for Government Insights at Deloitte, says the push for e-governance started during the dotcom era, however, it didn’t take off in a truly transformational way till about a decade later.

The author of nine books on the public sector, spoke to Fortune India about how digital technologies have the ability to fundamentally transform the way the sector operates and the challenges that still remain. He says, in the next decade, both public and private organisations will systematically apply artificial intelligence (AI) to their operations—a move which will change how work gets done.

Edited excerpts:

Tell us about the digital transformation of the public sector over the years. What are the challenges that you still see?

During the dotcom era we had a big e-government push, that was 15-17 years ago. For the most part what governments did was: it was more like a Hollywood storefront, they would put up a website, they didn’t change their back-end systems, it wasn’t great customer service, things were better, but it wasn’t transformational. In many respects, we didn’t always have the technology, we didn’t have the talent, people weren’t used to having a great digital experience even in the private sector then. Then starting around 2006 or so, we started seeing this push towards, what I call, the second wave of digital government, which was really about basing it on more modern digital practices that you had seen in the private sector.

So we are talking about over a decade of what we call modern digital with some of the governments, with the U.K. being one of the larger countries, to lead the way. Earlier, we would do these giant technology projects that were custom and very expensive. Now it’s about doing smaller projects, using agile development, having a user focus and bringing in people who understand user experience, and user interface and a whole variety of different people who never really worked with government before—the kind of people that are developing apps and websites in the private sector.

A lot of governments still haven’t created those sort of capabilities, they might be doing apps but they are still not doing digital in a modern way. One of the questions I ask governments is – how many people do you have who have experience in doing user experience and user design. In some countries I get a blank look. Those are some of the things to look at.

The government doesn’t need to reinvent the wheel, the commercial sector has been doing this for over a decade and it’s just about bringing in a lot of those capabilities and applying them in the government. Globally, the next stage of digital transformation of governments is AI.

Will you break it down for me, How will AI in governance work?

There are a lot of different types of artificial Intelligence. Take something like speech recognition, for example chatbots that are able to understand you. Then there is speech/language machine translation or speech translation. We do projects around the world where we used to have translators but we don’t have to have them anymore. We have also got computer vision, which is facial recognition, biometrics. Computers are better now than humans at recognising faces, and better than humans at recognising human emotions. You can use that for security purposes, at airports, for investigations, or forensics.

There is a lot going on right now, there is a lot of experimentation. Very few organisations, public or private, have systematically applied this to all of their operations. That is what we are going to see over the next decade or two, everything will become smarter, and it will change how work gets done.

When companies speak of digital transformation, It is usually done it in stages, does that happen in the public sector as well. What about countries like India and China with a large population?

What’s interesting is that when you are looking at digital and AI together, having that many people doing digital transformation in some respects can be difficult because the systems have to be there. But when you look at it from an analytics perspective and AI, it’s all about training sets and having a lot of training set data. One of the reasons why China is starting to catch up to the U.S. in a lot of ways is because of the massive amount of training set data that they have.

The countries that were leading in digital were the smaller ones like Estonia, New Zealand, and the U.A.E. because they could do it from a central perspective, they could create that citizen experience, it didn’t cost as much. But from an AI perspective, their training set data is not half as much valuable.

Developing countries have also leapfrogged in terms of technology, especially with mobile. So all these big, legacy, expensive, custom-developed systems, a lot of countries have had the opportunity to skip past them and save billions and billions of dollars of cost.

In India, roughly about 40% of people use smartphones and have access to the Internet, so in such a situation how do you transform the public sector when people don’t have access to digital?

The adoption of mobile technology is the fastest of any technology in human history. Especially when we look at the cost, it has gone down. We have had a billion people that have got access to the Internet in the last few years, and another billion is going to come on in the next few years.

There are a lot of providers that provide very low-cost, kind of quasi-smartphones in India. You go to any village in India or Africa, you will find mobile technology. A lot of providers who have become really good at creating apps for people who don’t have sophisticated smartphones. There has been a lot of innovation in that area. So, yes it is going to take some time for 100% penetration, but look at every other major technology in history, their adoption was so much slower.

This brings us to the question of regulation. We see technology and regulation are often at loggerheads. Do you think it is difficult for regulation to catch up with developments in technology?

It is important to ask how do you regulate emerging technology, technologies that are advancing at an exponential rate. Whether it is AI, or drones, autonomous vehicles or fintech, the argument is that you can’t regulate them in a traditional way. The first problem is the pacing problem, where technology is advancing rapidly and policy is catching up from behind. The policy lifecycle is 10-15 years, and a year is a lifetime when we look at AI and how that’s developing.

You don’t want to regulate too fast or too slow. What we talk about is a new set of principles for looking at these emerging technologies, things that are agile, adaptive regulation, government sandboxes where regulators are working with startups, providing space to test technologies, and better understanding technologies and watching and learning about technologies and getting close to them. Regulators all over the world are trying not to regulate too quickly.

The other issue is the notion of “innovation arbitrage”. If you have very restrictive regulations on say drones in a country, all the people who are trying to innovate, manufacture drones and test them, they will not work in that country, they will go outside. And it’s happened, in fintech, in autonomous vehicles. You have to create a climate that is competitive. Regulators have to be much more agile than they were ever before.