This year’s 21st CEO Global Survey from PwC highlighted growing anxiety among the world’s CEOs about the speed of technology-related change. It was notable as one area of concern from company decision makers that wasn’t focused on things that are happening outside the organisation (and its control).

PwC forecasts that AI will contribute an additional US$15.7 trillion to global GDP by 2030, an increase of 14.8%. But, while this will be a “boon to the overall economy… [it] will come at great cost to those who cannot rise to its challenges in time.”

Part of the anxiety no doubt comes from the fact that for many of us, artificial intelligence is a complicated ‘black box’. As PwC’s 2018 AI Predictions report explains, “many AI algorithms are beyond human comprehension”, and others are shrouded in the secrecy of intellectual property. So it’s no wonder that for executives AI is a complicated concept to wrap a brain around.

The below infographic from Futurism gives a gratifyingly simple explanation of machine learning, which is the type of AI that learns without human intervention.

Essentially, machine learning involves feeding a program enough data so that it can learn how to do something via a mathematical model. For instance, to recognise that an image is of a dog and not a cat.

There are a lot of different types of models – those that predict based on patterns, those that mimic the way human brains work, those that use probabilities and so on. But the end goal of all will be that the machine will learn – no matter what kind of model it’s using – to do the task on its own (to varying degrees).

More often than ever, terms like AI, machine learning, deep learning, artificial neural networks and bayesian networks are being used interchangeably. And while they aren’t the same, they are all variations on a theme.

Generally speaking, AI is the overarching concept of a machine doing things in a smart way without following ‘dumb’ rules (like a calculator would). Machine learning is the way in which it can become intelligent without humans giving it those rules. The others are just different ways, or models, that allow the robot brain to get from simple to sophisticated.

Easy, right?

Just kidding! Don’t worry, the black box that is your brain will understand it eventually, just like the robots will.


With thanks to Phil Bolton, a director in PwC Australia’s Experience and Insights team.

Source: Futurism