Artificial intelligence and the future of work: Will AI replace our jobs?

As A.I. technologies rapidly emerge, their impact on the job market raises many questions. Although some might fear a looming “jobs apocalypse,” the International Labour Organization (ILO) gives a more nuanced perspective. Certain tasks are more difficult to automate, and new technologies create new opportunities.

UNRIC spoke to Janine Berg, Senior Economist at the ILO, on how A.I. might affect the world of work, including job creation and displacement, as well as its impact on regions and genders. This article explores the potential dangers and advantages that A.I. brings to the workplace and how best to prepare the workforce for an AI-driven future.

How will A.I. impact the job market in the coming years? What opportunities and challenges will A.I. bring in terms of job creation and displacement?

It’s very difficult to predict what will happen in the future, not least because technology continues to evolve. ILO research does not suggest that there will be a “jobs apocalypse”. There are many tasks that A.I. cannot replace or where it would not be cost-effective to replace a human. Many occupations deemed “essential” during the COVID-19 pandemic, such as emergency medical technicians or food preparers, involve jobs that cannot be automated.

ILO research on the effects of generative A.I. on the labour market finds that at most 2.3% of jobs across the world have the potential to be fully automated. Moreover, this 2.3% does not account for the many jobs that will be created thanks to new technology. Twenty years ago, there were no social media managers; thirty years ago, there were few web designers. No amount of data modelling would have predicted the vast array of occupations that have emerged in recent decades. Just think of the video rental stores that existed 25 years ago and how they have given way to cell phone repair stores.

We should also remember that many systems that appear automated are not, with human labour working in the background, often under poor conditions. This is true of A.I., which relies on millions of human labourers to train and verify data. In other cases, paid staff are replaced by the unpaid labour of customers, for example, through self-checkout machines in grocery stores.

Will the impact of A.I. on jobs be balanced between different regions and genders?

While the potential for job displacement from A.I. is not huge, it is concentrated. Our analysis finds that clerical support workers are most exposed to generative A.I. technology, with 58% of tasks in this occupation having medium-level exposure and 24% having high-level exposure. This stands in contrast to other occupational groups, in which the largest share of highly exposed tasks oscillates between 1% and 4%, and medium-exposed tasks do not exceed 25%. As women are over-represented in clerical jobs, they are more affected by the potential effects of automation. According to our analysis, the potential impact on women is roughly 2.5 times greater than on men. We estimate that 1.4% of men’s jobs globally have the potential to be automated, compared with 3.7% of women’s employment. In high-income countries, where clerical jobs make up a greater proportion of occupations, the effects are more striking: 2.9% of men’s jobs compared with 7.8% of women’s jobs.

Another imbalance concerns the potential benefits of A.I. between rich and poor countries. ILO research indicates that many occupations could benefit from A.I., for example, by replacing mundane tasks and freeing time for more creative and interpersonal work. But these productivity gains depend on having the physical infrastructure and digital skills in place, which many parts of the developing world currently lack.

In a recent ILO study by my colleague, Pawel Gmyrek, co-authored with Herman Winkler of the World Bank, they assessed how many jobs that could be “augmented” through generative A.I. currently use a computer at work.  They found that in Latin America, nearly half of the occupations that could potentially benefit from A.I. augmentation do not use a computer at work. In contrast, jobs that are at risk of automation tend to already use computers, suggesting that the potential negative effects may appear sooner.

How can governments and organizations prepare the workforce for the changes brought by A.I. technologies?

It benefits both business and workers when employees are engaged in the process of technological integration. Workers know their jobs best and can provide insight and feedback on which tools could better support their work, which tasks could be automated, and what training is needed. Problems often arise when technology is deployed without worker involvement, both in terms of the technology not working well and because it diminishes workers’ sense of purpose and commitment to the firm.

The best approach is to integrate technology through social dialogue between employers and workers, and their representatives. This can occur through formal settings, such as works councils or collective bargaining agreements, or more informally, in workplaces with a high degree of employee engagement, such as organizations that support teamwork, problem-solving and decentralized decision-making. Studies on Europe have shown that it is the countries with stronger and more cooperative forms of workplace consultation, essentially the Nordic countries, followed by Germany, where workers are more open to technological adoption.

Social dialogue is essential to manage all the implications that new technologies, including A.I., can bring about in workplaces and the labour market more broadly.  It is crucial to mitigate technological unemployment, as it encourages redeployment and training over job loss, as well as for integrating technology into the workplace, to ensure that it is done in a way that benefits both firm productivity and job quality. Social dialogue is also vital for designing skills programmes that can train workers for new career opportunities, and for ensuring that the skills programmes undertaken are well-suited to the demands of the labour market.

What policies or measures should be put in place to ensure a smooth transition in the labour market as A.I. advances?

I already mentioned that social dialogue is key to managing the transition. Another concern is the global A.I. or digital divide. We need to ensure that the many dimensions of A.I. governance are addressed through international governance. In our recent joint report with the UN Secretary-General’s Envoy on Technology, Mind the A.I. Divide, we proposed three key policy pillars to address these challenges and harness the development potential of A.I.

First, strengthening international cooperation on A.I. is crucial. To close the A.I. divide, we must foster a global network for knowledge-sharing and establish joint training initiatives. By building collaborative research partnerships and ensuring equitable access to A.I. resources, we can help all nations, particularly developing ones, to benefit from A.I. advancements fully.

Second, building national A.I. capacity is essential. Countries need to make significant investments in education and digital infrastructure. This includes integrating A.I. and data science into educational curricula and ensuring that A.I. tools are widely accessible. National policies should promote human-centred A.I. development, ensuring privacy, safety, and workers’ rights are protected.

Third, integrating A.I. positively into the world of work is vital. This involves ensuring that A.I. contributes to creating decent work opportunities and supporting workers’ reskilling efforts. It also requires addressing gender- and youth-specific challenges. Crucially, social dialogue should play a key role in managing A.I.’s integration into the workplace, ensuring that the voices of workers are heard, and that A.I. enhances rather than diminishes job quality.

Further reading: