Which Impact Will AI Have on Job Automation?

A significant part of the workforce could be automated via AI, given its potential to replace repetitive and/or easily automatable tasks, job losses and a refocusing of the workforce around new jobs. Several studies estimate the number of jobs at risk around the world: for the OECD, 14% of jobs worldwide are at risk of automation. PwC estimates that 326 million jobs are at risk, close to Goldman Sachs' estimate of 300 million jobs. A McKinsey study points out that it will be activities rather than jobs that will be automated, suggesting that 45% of activities in the US (representing $2 trillion in wages) can already be automated with existing technologies.

This will likely depend on two factors in particular: the proportion of time an activity is spent performing repetitive tasks and the level of training and qualifications required to perform the activity. Sectors such as transport, logistics, manufacturing, retail sales, fast-food kitchen staff, or secretaries and counter staff are among the jobs most at risk according to the OECD.

Estimates of the share of jobs at risk of automation by selected activity, global average, according to the OECD

Nevertheless, publications by French authors Nicolas Bouzou and Laurent Alexandre provide reassurance about the future of jobs. Indeed, major scientific discoveries in history have always been accompanied by fears of job losses, with no prospects for the future: for example, printing threatened the work of monks who copied manuscripts, the automobile threatened the jobs of horse breeders and roadside workers, etc. Each time, new and unknown jobs appeared, and the boost in productivity generated by the discovery led to a major social advance for the population. During this AI revolution, the authors insist on the empathetic role of many human employees, which the machine cannot match or replace: a reassuring nurse, a fine psychologist salesperson, a financial advisor taking into account the family environment of his client, etc. According to Tesdal Neely, professor at Harvard Business School, two thirds of today's jobs did not yet exist in 1940, and "each technological revolution has created more jobs than it has destroyed". Therefore, by studying the history of previous technological revolutions, we can reasonably conclude that while easily automatable jobs will likely disappear, many affect-related jobs should remain, and a significant number of unknown jobs will appear.