By using our site you accept the terms of our cookie policy


What’s on the cards for AI?

Authored by K.R. Sanjiv, Chief Technology Officer, Wipro Limited

If you have ever played online poker you will know that there are millions of different potential situations. If a human played one situation every second they would need several million years to learn them all. The game is sophisticated, complex, requires considerable strategy, and like all poker games it calls for the innately human ability to bluff – and yet, last year poker playing AI beat four top human professional players.

Bridges, a supercomputer owned by Libratus, demonstrated that humans aren’t the only ones who can bluff, take a chance and win with unerring regularity. What Libratus achieved is spellbinding. Its creators, Tuomas Sandholm and Noam Brown of Carnegie Mellon University, did not provide the program with a complete set of situations, but rather used a small sample as Libratus played against itself for days, and accumulated a complete library of strategies through trial and error.

But what do developments like this mean to the real world? AI, Machine Learning and Data Analytics have fast become integral to every industry, and are soon to change even more. Yes, poker may turn its focus to classic face-to-face games, where bots can’t tilt the playing field. Who wants to lose every hand to a poker player that knows it all? But when AI works in tandem with humans as a team, it gives way to an entirely new set of possibilities. According to Gartner, by 2020, 20% of citizens in developed nations will use AI assistants to help them with an array of everyday, operational tasks. By 2021, AI augmentation is set to generate $2.9 trillion in business value and recover 6.2 billion hours of worker productivity.

How could this be applied to other industries? Let’s take the example of investors and traders playing the stock markets. Typically, they will synthesise information regarding their order books, listed companies, look at factors such as cash flow, interest rates, future plans, regulatory environments, market trends, currency behaviour and competitive signals, to make a decision. Humans can do all of this – and they spend endless hours studying these factors – but now a computer can beat them in terms of speed and accuracy.

Imagine an algorithm that can be trained to automatically trade, in say bitcoins, by self-learning. Now, imagine that the algorithm or trading engine is hosted on cloud infrastructure and you can connect it to your bitcoin account using a public API. You could be making money 24x7, even while you’re sleeping.

In the real world, financial advisers are already using bots to improve decision-making. Morgan Stanley, for example, has announced that it will arm its 16,000 financial advisers with machine learning algorithms that will take over routine tasks. Morgan Stanley forecasts that robo advisers will manage $6.5 trillion of wealth by 2025, or 5% of global wealth.

Interestingly, what Morgan Stanley is doing in the world of investment is a little different from what Libratus did for poker. Unlike the poker playing AI, Morgan Stanley hopes to put AI at the disposal of human advisers to improve outcomes and service levels for their wealthy customers.

But with the onset of AI showing even greater ability to take on human complexities, many suggest that Singularity (the point at which the distinction between humans and machines will be erased that is) could be just around the corner. Not only are compute capabilities expanding exponentially, but costs are coming down. Robots are getting smarter, for sure. Lego makes millions of bricks every hour and packages them with absolute precision using robotics; robo-assisted surgery is helping make complex procedures safe and cheap; beautiful homes are being built (literally) overnight by a bunch of automated arms driven by CAD drawings. But are we close to being run over by machines? Not at all. Firstly, the ability to do everything requires more than brute compute capability. It requires the exchange of sophisticated and fine-grained knowledge between machines. But machines don’t have a uniform way of collaborating between them. For machines to create a complete sum of all parts is decades away.

Admittedly, we have technologies like Machine Learning, Neural Networks, Natural Language Processing, Cognitive Computing, but there are components that no one is even working on to bring singularity closer. The truth is that we don’t even know all the critical factors and conditions required for such a level of connectedness, collaboration and knowledge exchange. For example, how do you create AI engines that can invent and create new innovations which are contrary to what it has historically learnt from past experiences? Industries from manufacturing to medicine are at the cutting edge of technology, busy re-writing business models and human life cycles. But the path to a completely unified AI and human experience has many stops and questions to answer, while man and machine continue to develop a partnership.