Amazon boss Jeff Bezos doesn't think the robots are going to take all our jobs or stalk the human race with malign intent.
“The idea that there is going to be a general AI overlord that subjugates us or kills us all, I think, is not something to worry about. I think that is overhyped,” he says. Bezos takes this view because, as of now, he says, we only have “narrow” AI and that we are nowhere near being able to build “general” AI that could operate autonomously.
Bezos is not alone. There's now a groundswell of opinion that suggests the robots are somewhat more benign than we've been led to believe. Yes, in many cases they're going to replace humans and up to 66 million jobs are at risk within the OECD due to automation. But a 2018 report from the Paris-based economic think tank suggests that the risk of a robotic clean sweep is lower than previously forecast.
It also depends on where you are. If you live in Norway, for example, only 6 per cent of jobs are ripe for automation compared with 33 per cent in Slovakia. The figure for Ireland is around 15 per cent, and 12 per cent for the UK.
The report says that, on average, about 14 per cent of jobs in OECD countries are “highly automatable” while a further 32 per cent are likely to experience significant change in how they are carried out due to technologies such as robotics and AI.
Entirely redundant
“Some jobs may become entirely redundant although the extent of the automation will likely depend on policy, institutions and social preferences,” the report says.
We shouldn’t be overly surprised. If history has taught us anything, it’s that calamities with star billing don’t always turn out as expected. When farm workers were displaced by combine harvesters, they found jobs in factories. When factory workers were displaced by automated production lines, they found jobs in services. Economies and people have always shown resilience and the capacity to regenerate.
US economist Oren Cass, author of The Once and Future Worker, takes issue with four of the assumptions germane to the "robots are coming" theory. He says:
– it miscalculates the pace of the change;
– it is predicated on automation only taking jobs, not creating them;
– it assumes robots can do everything humans can;
and
– it overestimates how clever 21st century humans actually are.
It is now generally accepted that estimates of how many jobs robots will replace are subject to major uncertainty. In addition, while some jobs will go, new ones will emerge so the total of people in work may continue to rise.
The benefits of introducing more technology into a tight labour market can also help improve productivity leaving scope for wage increases without affecting profitability.
“Engineering bottlenecks”
At this point, there are still barriers separating us from the robots, most notably in the shape of what are known as “engineering bottlenecks”. These are tasks that cannot be easily automated because it’s a challenge to write the code. Tasks that fall into this category include those requiring perception and manipulation in complex situations and in tight spaces, anything involving original thinking and creativity and tasks that require social intelligence and/or a caring attitude.
However, the OECD aims a shot across the bows of those individuals and organisations with at-risk jobs who are not reacting to the threat. In particular, it says training is not working to offset the risks because “participation in training is significantly lower for workers in jobs at high risk of automation than for other workers”.
If a job can be easily explained then it’s ripe for automation and jobs considered at highest risk are those in food preparation, cleaning, labouring in mining, construction, agriculture and manufacturing, and services such as postal/courier deliveries and, of course, land transport with the advent of autonomous vehicles.
Young people are most vulnerable followed by older age groups while those without training or qualifications will be very exposed.
The good news is that we’re not heading towards a jobless society of couch potatoes hooked on daytime TV. Every second job in the OECD could, in theory, be automated – but it is only a possibility.
Job automation
“The actual implementation of full or partial job automation will depend on a range of other factors such as technology penetration and adoption, the cost of human labour relative to the new technologies, and the social preferences for automating certain tasks,” the report says. “Take-up of new technologies lags behind their technical feasibility and some innovations are never widely implemented.”
The other good news is that the risk of automation declines as education and skill levels rise. However, that's not a good enough reason to be complacent, especially if you work in the professions, according to Prof Richard Susskind. He's the co-author of The Future of the Professions, which goes into some detail about the scenarios facing the professions as a consequence of advances in technology.
Susskind has spent the last 30 years analysing the impact of technology on the legal profession in particular, and has written copiously about it in books such as The End of Lawyers. He predicts a future world of online courts, AI-based global legal businesses, and internet-based simulated practices.
He also says that the threat to the legal profession as we know it doesn’t come from lookalike competitors, but from the 3,000-odd legal tech startups that want to do for the law what Amazon did for selling books.
