Why your job is safe from AI (if you happen to be a human)
A Japanese insurance firm recently announced it will be replacing more than 30 employees with an artificial intelligence (AI) system that can accurately calculate payouts to policyholders.
The firm claims the move will increase productivity by 30% and save around £1 million every year –numbers that will likely make other organisations sit up and listen if they turn out to be true.
Should us humble human employees be worried, then? Is it really only a matter of time before we’re all replaced by robots?
The short answer is: no. Absolutely not.
But before I give you the long answer we first need to clarify what AI actually is.
This might seem odd – we’ve all been talking about AI for years in some capacity (I should know – I first studied it at college back in the mid-80s!). But here’s the thing: AI means different things to different people. If you take a large amount of data and predict it’s going to snow tomorrow, that’s not AI. That’s insight. Predictive, yes, but hardly intelligent. Yet many seem to refer to this as AI.
To me, AI is when you can take seemingly unrelated sets of data and mash them together to create deeper insights than you otherwise could.
That’s what the human brain does. Something might look completely irrelevant on the surface, but the brain can spot a pattern or make a connection, come up with a hypothesis and investigate.
With that in mind, here’s why we don’t need to be afraid of losing our livelihood to machines…
You’ll always need human intervention, input and intelligence
The conversation around AI in the workplace always seems to revolve around whether this technology will replace human beings altogether. But that assumes future AI will have the ability to automatically act upon any insight it uncovers.
The thing about humans is we don’t just figure things out – we also work out what to do with that insight once we have it. Predictions deal with what might happen and why, but then it’s up to people to know how to use that information.
AI should not be built to replace us but rather to mimic us. In the decades following the early discoveries in AI the world has largely drawn a blank when it comes to making use of it. Why? Because we were trying to make the AI be the human mind rather than supplement it. We should have been using AI to support our own thinking and enable us to do jobs that are much more complex. That’s where we seem to be headed now.
Reports of workplace AI focus too much on redundancy and not enough on the positive impact this technology can have. Think of it as a high-speed, high-volume neural network that provides us with a set of insights from which we can derive answers using the human and non-deterministic part of our brain.
In future AI will bring everything to us – all the data and insight we need – and we will apply our own intelligence to determine whether the AI has done the right thing and then approve or reject the outcome accordingly. But that validation requirement will always be there. The AI will become increasingly intelligent over time and we will be pushed to go deeper and deeper in our own analysis, but the human element will remain.
And, if anything, this will likely create more jobs as more people are needed to carry out the analysis and action based on the insight the AI provides.
AI is what we make it
Let’s face it: AI is only ever as good as the models and information upon which it’s built. It is powered by data, and humans are in control of that data and the way the AI goes about processing it. We point the AI at the data sources we want it to analyse and we build the models that enable it to traverse those sources and come to a determination.
The beauty of the human brain, however, is that it can spot potential sources of information on the fly. And it doesn’t need to be trained on a new model to accomplish that.
Whether AI one day becomes advanced enough to achieve that human element on its own remains to be seen, but it’s much more likely we’ll see an increasing number of use cases designed to support and enhance what people do in the workplace rather than replace people altogether.
For AI to be exploited and succeed (and it will succeed), we need to focus on discovering and developing use cases that support business but also our wider society.
When Big Data became mainstream years ago, people said it would change the world. And it did. But the biggest problem people had to deal with was working out what to do with all the data they now had access to. Where do you even start? It was only when people started to discover ways to break down and apply all that information that Big Data really took off.
We create our own destiny
Maybe at some point AI will be able to create use cases itself, which in turn will feed into an AI engine to create a classic circular reference situation.
But even then, the way we apply any insight gained will (and should) always be down to us.
Back to that original question, then: should we be worried that AI might one day take our place at work?
No. Humans are here to stay and it’s our job to exploit these technologies for the good of the world.
AI could have a negative impact, yes. But we’re in control of that. We have to embrace this change and then channel it in the right direction.
Check out the below videos from Fujitsu Forum 2016 to learn more about how digital technology is transforming the workplace and our EMEIA CTO’s view on AI: