Imagine you’re a chief executive and someone offers you an idea to simultaneously cuts costs and boost productivity. What’s more, they’re not small numbers – you could reduce overheads by 13% while boosting efficiency by nearly the same amount. It’s too good to ignore.
This scenario isn’t fictional and is the opportunity facing businesses around the world as a range of new technologies mature. In the above example, these numbers are from the mining industry in Australia where driverless trucks have now been in service for several years.
Impressive as they are, these figures pale in comparison to some situations, such as a Chinese factory that reportedly replaced 90% of its workforce two years ago and increased productivity by a staggering 250%.
The technology that allows Chinese factories to automate knows no borders. To illustrate, it’s possible to buy advanced industrial robots in New Zealand and easily deploy them on a factory floor. One of the better-known examples is called Baxter, which can be up and running in a day with virtually no programming for around $30,000.
Given the readership of the NBR, many people reading this column are probably seeking comfort in the illusion that your job can’t be easily automated. You would be wrong. For example, over the past 17 years, automation has meant one trading desk on Wall Street has reduced in size from 600 traders down to just two.
In the face of these examples, people are seeking out professions that offer more certainty and programming is touted as one of these. However, while demand for technology skills is high this, too, is a profession that’s going to change quickly.
Earlier this year researchers at Google created a piece of artificial intelligence software that was designed to create another piece of artificial intelligence software – in effect, smart software designed to build other smart software. What they found was the resulting software – built by software – performed the desired tasks better than an artificial intelligence that had been built by humans.
If you haven’t quite grasped the implications of this, stop and re-read the above paragraph. Even though this is a piece of laboratory research, it demonstrates that computers now can program themselves better than people can.
The implications of this led the influential technology magazine Wired to suggest, in the future, the idea of programming will become a job that’s similar to training a dog. Nobody will need to actually write the code but people will need to guide the software so that it performs optimally.
The above examples are the tip of the iceberg.
In the US, long-distance truck driver is one of the most common occupations in many states. It is also one of the jobs that’s highly likely to be automated in the very near future as autonomous delivery trucks are perfected. That’s not far away and last year a division of Uber performed the first automated long distance delivery of goods.
There are many competing arguments about the speed and impact of technology on jobs but one thing is for certain – the nature of work is about to undergo a massive change.
Machines creating machines
While some people argue the world has seen similar changes in history, such as the first industrial revolution, what is fundamentally different now is that machines are creating machines that don’t just perform better but also negate the need for humans to be involved at all.
It’s highly likely that you’ve already experienced this effect many times in your life without realising it. If you’ve travelled in a commercial aircraft in the past decade, you’ve put your life in the hands in one of the world’s biggest robots.
Once a commercial airliner takes off, most of the flight is handled by the software we commonly call the "autopilot," which is only usually switched off in the last few minutes before landing. In many airports now – including Auckland – the plane can land itself when the human pilots don’t have enough information to perform the task themselves.
This points to a new way of working that is likely to become increasingly common – trained people working with machines to augment their decision making. In the near term, this type of collaboration is unlikely to result in wholesale job losses, with a McKinsey report released earlier this year suggesting less than 5% of jobs are able to be fully automated.
However, there are more sobering predictions, chief among these being a now famous Oxford University study published in 2013 that predicted the likelihood of various jobs being taken over by machines in the next decade.
The research found that professions such as telemarketers and insurance underwriters have almost a 100% chance of automation, while many healthcare workers – such as recreational therapists – have an extremely low chance of automation. A closer examination of these predictions reveals a repeating pattern – jobs that are primarily rules-based are prime candidates for automation, while tasks that require creative and social skills are unlikely to be automated.
While there are widely varying reports about the impact of technology on jobs, one thing is clear – it won’t be long before philosophers start asking new questions about the purpose of life and the need to work.