This report (the fourth in our series on artificial intelligence) outlines research on the impact of AI—encompassing machine learning and robotic intelligence—on employment, largely within a five-year timeframe. It draws on data from the USA, Europe, Australia, and other nations.

The degree of job creation or loss caused by AI’s entrance into the workplace is contested. Some authors anticipate high unemployment within five years, while others are less convinced, viewing AI as an assistant to human labour rather than a replacement, with potential to increase national productivity. Literature also raises the possibility of job creation rather than loss. Caution with prediction relates to the capability, utility, and cost of the technology, as well as its ease of use and human sentiment.

Despite this, we can make provisional projections. Accurate longitudinal research has broken down the work people do into tasks (20,000 plus) across all occupations and industries. These tasks are matched against AI’s current and emerging performance capabilities. The findings enable analysts to cautiously chart the impact of AI’s adoption on the workplace.

The picture is very bleak, with potentially 40-60% of secretarial, administrative, and customer service roles exposed to and replaced by AI today. Many other occupations will experience similar threat levels over the next five to ten years. High-earning professionals are not exempt from risk. Introducing AI into our workplace will also change the social status of employees, flattening it for some while increasing it for others.

Just because AI can perform human tasks does not mean it will. Important variables will either accentuate or moderate its impact on our workforce. These include the area of work, gender, age, ethnicity, and specifically, whether employers regard AI as a replacement or assistant technology. Women in a number of industries and occupations are more vulnerable to displacement than men. Some ethnicities, too, tend to cluster in certain fields of work, such as manufacturing or finance, which are at higher risk of retrenchment.

This report anticipates a future with higher-than-current levels of unemployment. It expects new jobs to be created in relation to machine learning and artificial intelligence, but not as many as are displaced. It raises questions about the value, importance, and need for work. It recommends planning and preparation for a future where, even though jobs might be fewer and competition for employment fiercer, as a people and a nation we will continue to prosper.

Dr Paul Henderson on Seven Sharp