PwC’s Anand Rao: ‘We are only in 1984 in terms of the evolution of AI’

13 Apr 2018

Anand Rao, PwC global leader for artificial intelligence. Image: PwC

AI will augment people’s capabilities but won’t take all jobs. However, there will be socioeconomic upheaval, warns PwC’s authority on AI, Anand Rao.

Anand Rao is PwC global leader for artificial intelligence (AI) and is the consulting giant’s innovation lead for the US analytics practice. Rao has 24 years of industry and consulting experience, helping senior executives to structure, solve and manage critical issues facing their organisations.

He has worked extensively on business, technology and analytics issues across a wide range of industry sectors including financial services, healthcare, telecommunications, aerospace and defence, across US, Europe, Asia and Australia.

‘Open source innovation is actually accelerating the growth of AI’
– ANAND RAO

His work has included behavioural economics; simulation modelling; global growth strategies, marketing, sales and distribution strategies; online, mobile social media strategies; customer experience; multi-channel integration; risk management and compliance; large-scale programme mobilisation; and management. Before his consulting career, Rao was the chief research scientist at the Australian Artificial Intelligence Institute, a boutique research and software house.

Rao has also co-edited four books on intelligent agents and has published more than 50 papers in computer science and AI in major journals, conferences and workshops. He is a frequent speaker at a number of industry and academic forums.

Rao was awarded the Most Influential Paper Award for the Decade in 2007 from the Autonomous Agents and Multi-Agent Systems organisation. He holds an MSc in computer science from Birla Institute of Technology and Science in India and a PhD in computer science from the University of Sydney, where he was awarded the University Postgraduate Research Award. He also received an MBA (distinction) from Melbourne Business School.

What are some of the main responsibilities of your role?

I lead our global AI innovation effort in data and analytics. The way I got into the role was, I started my PhD in AI from Australia in 1985 and subsequent to that, I was working in an R&D organisation in AI security. We were primarily working for the US defence sector and NASA on the aerospace side and on a number of defence-related projects in Australia. I did that for 12 years after my PhD and then I shifted more into management consulting in the late 1990s and, since then, I’ve been doing management consulting in Australia, UK and Europe.

I moved to the US in 2006 and that’s when we started some of the core analytics at PwC – international language processing, machine learning and, more recently, deep-learning agents.

Where are we at on the AI journey relative to the evolution of computing?

It has been around for a long time. The term was coined in 1956, and there have been various ups and downs in AI over the last 10 years – it has picked up substantially.

If you look at what has propelled AI into its recent incarnation, there have been two other big phases of AI. There has been a lot of hype and overblown expectation and, when AI has not been able to meet those expectations, funding was withdrawn. That happened twice before and we are now in the third AI spring. Everyone will say this time it is different, and it is different. But we will see.

What has propelled AI in this era or recent phase? Firstly, the compute power is working its way to a stage where it is really powerful. And, one of the key breakthroughs is that power is exponentially increasing – serendipitous for AI in the GPU and how that architecture precipitated deep learning, and that, essentially, was one of the sparks for the great excitement in this period.

Things that could not be done in the past due to computational limitations are now possible – not only with the speed but also, the GPU is massively powerful.

The reason why machine learning has become more mainstream, compared with the 1980s and 1990s, is more around the data. It really has become embedded in everything because no one needs to talk about big data, because there is big data wherever you go and people are using that.

And AI is the next incarnation of that in the sense that deep learning needs that data; it is hungry for that data and it is eating that data, making AI more viable.

PwC’s Anand Rao: ‘We are only in 1984 in terms of the evolution of AI’

From left: Richard Day, partner, PwC; Anand Rao, global leader of AI, PwC; Prof Barry Smyth, digital chair of computer science, UCD; and Enda McDonagh, partner, PwC. Image: PwC

How is open source software also contributing to the AI revolution?

There is also this whole movement around open source. If you look at the software development community in the 1990s versus today, it is fundamentally different. Software development in the 1990s meant writing line by line of code. Now, almost no one starts off with a fresh heap of code; you are essentially building on top of other people’s code and, in most cases, modifying it.

This is really accelerated and AI is contributing to that through open source algorithms. Open source has not only been in terms of code, but also in terms of data and sharing of that with the entire community. Open source innovation is actually accelerating the growth of AI.

The olden days of proprietary data and code are gone. Now, it is: make it open, get more people to use it and, that way, we’ll make more revenue.

What impact is AI making on productivity? Some believe AI will enable humans to be more productive while others believe AI will take jobs. Where do you sit on that?

My own view is that, definitely, AI will automate many (if not all) of the manual, cognitive, repetitive tasks, even non-repetitive tasks. All of this will be automated.

The early phase of that automation will continue. To me, when it comes to AI, we are in the early days of computer revolution. 1984 is when we had the first PCs and Macs – we are pretty much at that level. It is 1984 as it relates to AI adoption and we’ve come a long way in terms of the pervasiveness of computing.

This is not just in advanced countries but even more so in some of the developing countries that have smartphones. Now, in fact, people who are not literate in languages can speak more with people who can write a particular language, and the smartphones are becoming a conduit for them to know more about certain things. That is how pervasive that AI is becoming. It will take away all jobs but the AI will start helping us.

It’s not really man versus machine, it is always together with man and machine. That’s why we believe much more in the augmentation; we are going to be augmented with more and more intelligent tools.

In 1990, the amount of information available to me to write my thesis was my library and, in a very specific area I was working on, there were 20 books and that was all the information I had. But today, there is no excuse. Everyone expects you to know anything and everything on that topic because it is just one Google search away.

But how do you synthesise that data and apply it? We are essentially augmented with all the knowledge but the critical thing is, how do we use it? More insight? But what do you do with those insights and how do you apply it?

We don’t believe all the jobs will disappear. There is an economic and social angle. The jobs per se won’t disappear but certain job categories will diminish or disappear, and it becomes a social issue as to what happens to those people if administrative or manufacturing jobs go, what those people will do, and that is the issue.

I wouldn’t say AI is taking over everything. The social challenge is that, in the long term, people will adapt and, from a macrosocioeconomic perspective, everything will fix itself.

But the question is, are things moving so fast that there will be a large amount of displaced people? And how do we make sure that society doesn’t move in two polar directions, which would be detrimental to the overall growth?

This is already happening. The disparity between the top and bottom is bigger in US, not yet in Europe. In terms of automation and how the wealth is getting distributed, I think, irrespective of which path is chosen, things are moving fast.

The demand is for AI jobs but, for potentially millions of unemployed people, you cannot realistically say to them, go and learn Python and become a deep-learning expert – that’s not going to be possible. The younger people may but for ones that are displaced, it is not an option.

There are going to be real social and economic issues but, in a broader macroeconomic sense, I don’t see humans being replaced by machines or that we are going to be having so much spare time that we won’t be doing anything else but going to the beach and watching movies.

Want stories like this and more direct to your inbox? Sign up for Tech Trends, Silicon Republic’s weekly digest of need-to-know tech news.

John Kennedy is a journalist who served as editor of Silicon Republic for 17 years

editorial@siliconrepublic.com