The potential for rising unemployment is one of the main concerns internationally when it comes to the rapid growth of artificial intelligence (AI).
A report published last week by the Economic and Social Research Institute (ESRI) and the Department of Finance warned that as many as 200,000 workers in the State could soon lose their jobs as a result of the fast-developing technology.
Big tech companies have already slashed a significant number of jobs in the United States while stepping up investment in AI. Microsoft last year laid off 15,000 staff and Amazon has cut twice that number in the last six months. Both have a significant presence in Ireland.
Graduate roles in AI-affected industries are becoming harder to find. A report by recruiting group Morgan McKinley last January noted AI adoption is now a major theme across all sectors, but particularly within tech and accounting and finance, where it is “moderating” the demand for junior and generalist roles and “intensifying the hunt for experienced AI specialists”.
A September labour market update from jobs site Indeed noted a “stagnant market” for new graduates.
Though it sets out a number of potential outcomes, the ESRI and Department of Finance report concentrates on a “central scenario” in which AI causes 7 per cent of jobs in the State to be displaced. Currently, the State has an unemployment rate of just under 5 per cent.
Brendan Kelly, professor of psychiatry at Trinity College Dublin, says such a fall in the national employment level would affect the mental health of society in general.
“There’s a spillover effect into people who still retain their jobs … 7 per cent would be more than enough to impact on the mental health of everyone in society, not just those who have lost their jobs.”
A natural comparison can be made with the 2008 financial crisis, which brought about an unemployment rate of 15 per cent in 2010 and 2011.
“What we found during the economic crash of 2008 and 2009 was mental health in Ireland and elsewhere suffered substantially,” Kelly says. “The key driver of that was unemployment: people who lost their jobs rather than their finances.” He says unemployment is “uniquely corrosive” to a person’s mental health.
Many studies highlight that “being employed is good for mental health”, as it brings income security, self-worth and is an “important component of self-esteem for some people”, says Kelly.
ESRI associate professor Karina Doorley, a co-author of last week’s report, says Ireland could be better positioned than other countries to cope with a large-scale shift towards AI if the right preparations are made.
“We have a highly educated workforce, they should be able to upskill or they should already have transferable skills that can be used in other professions,” she says.
“The speed of that transition will be really crucial to how this plays out in the future.”
Where possible, Kelly suggests exposed jobs should be appropriately reshaped to the shifting AI environment, rather than abandoned in pursuit of an alternative career path.
Though it can be difficult mentally for people to reorient or retrain mid-career, he says it is better to be afforded that opportunity by a current employer as it avoids the “big bang of simply losing a job and becoming unemployed”.
Unlike previous technological revolutions that predominantly affected lower-skilled manual workers, AI poses the biggest threat to higher-skilled occupations, according to the ESRI research. Some of the most vulnerable sectors include accountancy, clerical work and telesales.
The potential for 7 per cent of jobs to be displaced is based on a study led by US economists Joseph Briggs and Devesh Kodnani.
They estimated that more than half of the tasks involved in 7 per cent of US jobs are exposed to automation and can therefore be replaced by AI. This figure, Doorley says, can be considered an “upper bound”, or maximum threshold, for the ESRI’s research.
Briggs and Kodnani suggest those who remain employed are likely to see wage increases of about 2.5 percentage points. These are broad estimates, and different occupations will be variously affected based on their vulnerability to, and capacity to benefit from, AI.
Doorley says there is “no consensus so far amongst international evidence” about what size the shock from AI will be. That is why the ESRI projected a range of other scenarios.
“In every previous technological revolution we’ve had, there’s been job destruction but there’s also been job creation. If you think about when robots were brought into the workplace, some routine jobs were lost but we needed people to build the robots, to service the robots and to write the code that made the robots work.”
On that basis, job creation is “very likely” within the AI revolution, Doorley says, but it was not something they were able to account for in their scenario analysis.
Doorley’s research forecasts that higher-income households would likely lose more as a proportion of their disposable income than lower-income households.
For example, the most exposed household in the “central scenario” is a middle to high-earning one that becomes unemployed due to AI.
“Some people are losing their jobs, while those that remain are seeing wage increases. That results in further polarisation in the income distribution,” she adds.
Kelly says research indicates that we would prefer to earn less if it meant “staying ahead of the neighbours in some way”.
[ The inconvenient truth about artificial intelligenceOpens in new window ]
“We are exquisitely sensitive to comparisons, either between households or maybe between people within the same household … Comparison matters more to us than absolute income.”
Kelly says it is a good time to remind ourselves that it is unhelpful to compare our lives with the situations of others.