Fears about automation eventually leading to mass unemployment have been a recurring theme for centuries. The Industrial Revolution introduced mechanisation and mass production. This created the first widespread fears of technological unemployment, inspiring the rise of the Luddites. Mass adoption of electricity triggered similar fears, as have periodic spikes in unemployment and the rise of robotics.
Today, these historic arguments are being played out against a new backdrop of artificial intelligence and smart automation. Cars can pilot themselves along motorways, virtual assistants access knowledge faster than humans ever could, and chatbots answer common queries without requiring any human input. Such developments have caused some observers to wonder whether certain types of jobs could disappear, following other obsolete roles like switchboard operators and film projectionists into the history books.
As short as it’s long
In the short term, those fears seem unfounded. The American technology sector grew by an estimated 56,000 jobs in the first half of 2019, maintaining a historic trend that new technologies often generate more jobs than they replace. Pessimism about the consequences of increased automation tends to be much higher in countries with struggling economies or high unemployment – a fatalistic ‘what’s going to hurt us next’ attitude.
By the 2030s, employment prospects are undeniably less rosy for people in low-skilled positions. The OECD estimates that 14 per cent of today’s jobs would be highly automatable through advances in AI and robotics. Menial jobs requiring little interpersonal negotiation or complex reasoning will be most affected – assembly line workers rather than repair technicians, for instance. Repetitive cleaning, mining, and manufacturing positions are seen as the most vulnerable to robotics and AI. For example, a robot could prepare cooking ingredients far faster than a commis chef, working around the clock unpaid while generating less food wastage.
A landmark report by PwC has divided the process of automation into three distinct time periods. An initial algorithm wave is already automating simple computational and mathematical tasks, in data-driven industries such as finance and insurance. This will be followed in the early-to-mid 2020s by an augmentation wave, where repetitive tasks like warehouse goods management and form-filling are automated. The third phase will be an autonomy wave, developing throughout the 2030s and involving driverless vehicles and water treatment algorithms.
Some industries will be impacted more than others and over different time periods. PwC believes that nine out of ten education jobs will remain safe long-term, whereas the majority of transportation and storage roles could be lost to partial or full automation. In the short-term, financial services and information management are threatened by improved data processing and predictive mapping tools. Manufacturing and construction may be vulnerable in the late2020s, yet literacy and interpersonal skills might always lie beyond the scope of emotionless algorithms.
Putting up barriers
It’s also worth noting that the potential for automation doesn’t mean its adoption becomes inevitable. Concerns about the safety of driverless vehicles (and the ethical dilemmas when an accident becomes inevitable) may prevent fully autonomous cars and HGVs ever populating our roads. Legal and regulatory constraints have held back certain advances in the past, while ethical concerns stalled cloning and GM crops. Social anxiety about job losses may act as a further brake on progress.
Although jobs created as a result of technological advances are likely to be more specialised than those being lost, research suggests that greater automation might not affect unemployment levels in developed economies. It seems the Luddites are still waiting to be proved right…