As A.I. technologies mature rapidly, people are asking what it means for the future of jobs. Some fear mass unemployment, but for a more tempered view we turn to the International Labour Organization (ILO). A.I. may shift how we work, but it will not replace everything — and could actually create new opportunities.
As part of a discussion with UNRIC, Janine Berg, Senior ILO Economist, reflected on how A.I. has the potential to affect the job market. She addressed the risks and rewards of automation on every region and demographic that could be most affected, and spoke about how to prepare for a future dominated more by A.I.
Will A.I. replace all of our jobs?
The short answer? Not likely. There are no signs of a “jobs apocalypse” coming according to research from the ILO. Sure, A.I. can replace some work and automate some tasks. But there are many things artificial intelligence cannot do — and jobs that involve hands-on care for other humans seems to be one of them. For example, roles like emergency medical technicians and food service workers were essential during the COVID-19 pandemic that machines couldn’t fill.
In reality, the proportion of jobs – 2.3% – that the ILO estimates are at risk of being entirely automated globally is small. And even that figure doesn’t account for the many new jobs that A.I. might help create. Just 20 or 30 years ago, look back and see, who could’ve predicted a job like social media manager or app developer? Remember video rental stores? Many of them have been replaced by businesses like phone repair shops and are a telling illustration of the way in which work is ever changing.
What all we don’t see behind the machines
It’s important to remember as well that a lot of systems we imagine to be “automated” are still, behind the scenes, dependent on human labor. A.I. depends on people to train it and confirm data by hand, often under trying or low-paid circumstances. In other instances, businesses simply shift the work to customers — such as self-checkout machines in grocery stores, where shoppers take on the task of scanning their own groceries.
Will A.I.’s Impact on Jobs Be Fair Across Regions and Genders?
The effect on jobs from A.I. is not huge one way or the other — but it is uneven. In our research, clerical support workers have the highest vulnerability to generative A.I. Some 58 percent of the tasks in these roles are of moderate A.I. exposure, and 24 percent of high A.I. exposure. Compare that with other occupations, where just 1% to 4% of tasks wind up in the high-risk range, and medium-risk tasks generally remain under 25%.
Now enter gender: because women are more likely than men to hold clerical jobs, they are more at risk from automation. Women actually bear about 2.5 times more impact than men. Three point seven percent of women’s jobs worldwide could be automated, compared with 1.4 percent of men’s. In richer countries the gap is wider: 7.8% of the jobs women do are at risk, compared with 2.9% for men.
Now let’s talk geography. There’s also a sharp cleavage between more developed and less developed countries. A.I. has the potential to increase productivity — things like automating mundane jobs so that people can do more creative and human-oriented work. But this works only if there is the right infrastructure and digital skills in place. Unfortunately these are still in short supply in many developing areas.
A new study from my colleague Pawel Gmyrek and the World Bank’s Herman Winkler delved into this. They discovered that in Latin America, close to half the jobs that could benefit from A.I. don’t even require a computer. Meanwhile, the jobs that are most likely to be automated are ones that are already heavily computerized — the downside may actually arrive before the upside.
How Can Governments and Organizations Prepare the Workforce for A.I. Changes?
Companies and workers need to collaborate if they are going to get the most out of A.I. technologies. It’s all about employees being a part of the process of implementing new tech, which is to everyone’s benefit. Why? Well, because nobody knows the job better than the worker. They can determine which tools would actually be helpful, which tasks they could automate and where to train.
“If workers are involved in the process, the tech and the humans can usually work together fine. The net result, apart from often creating tools that don’t work well in practice, is that the workers experience a sense of separateness and reduced motivation. Their belief in the mission and the organization is eroded.
The smartest way to manage the integration of A.I. and other technologies is through social dialogue — a two-way conversation between employers and employees (and their representatives). That can assume formal forms, such as works councils or collective bargaining agreements, or be more informal, for example at companies where teamwork and shared decision-making and problem-solving are part of the culture.
In Europe there is clear evidence for a relationship: in countries such as some of the Nordic nations or Germany with strong systems of workplace consultation and cooperation, new technology is more accepted by workers.
In other words, social dialogue is not just useful — it’s necessary. It will help manage the changes that A.I. brings, minimize the risks of job losses through redeployment and reskilling, and get technology to increase both productivity and job quality. It is also critical to inform skills programming that is relevant to actual labor market demands.
What policies or measures can help ensure a smooth labour market transition as A.I. evolves?
As A.I. continues to reshape the world of work, it is critical to prepare wisely — and that begins with robust social dialogue. Transparent dialogue among governments, employers and workers can help to navigate the shift. But another major worry is the worldwide digital divide. Not all countries are moving at the same ebb and that gap demands serious attention.”
In our recent joint report with the UN Secretary-General’s Envoy on Technology, Mind the A.I. Divide, we highlighted three policy pillars to help ensure a fair and inclusive transition:
Increase Global Cooperation on A.I.
We need countries to collaborate to bridge the A.I. gap. That would involve establishing global platforms for the sharing of knowledge, instituting joint training programs and forging research partnerships. When AI tools and resources are made broadly available, including to developing countries, everyone has a chance to grow and to benefit from technological progress.
Invest in National A.I. Skills and Infrastructure. We should invest in our nation’s A.I. skills and our infrastructure.
Each country will have to develop its own A.I. muscles. That includes serious doses of investment in education systems — bringing A.I. and data science into the classroom — and improving digital infrastructure. National policy should encourage people-first A.I. development, but ensure that privacy, safety and workers’ rights remain protected.
Make A.I. Work for Workers
A.I. is supposed to bring more good jobs, not fewer. Governments and employers must support re-skilling and take on the challenges facing women and young people in the workforce. All this would be centered around the most critical component: maintaining the centrality of social dialogue — ensuring that workers have a voice and that A.I. improves work rather than destroying it.