Recent advances in Artificial Intelligence have broadened the set of skills and abilities that can be replicated by automation technologies. In the past, computers and robots could only follow rules specified by programmers. However, machine learning algorithms, the branch of AI that has experienced the most important advances recently, can now make decisions without following pre-specified rules. Furthermore, AI technologies are able to work with unstructured environments and data. As a result, AI can help automate non-routine activities, contrary to computers that need codified environments and could only replace workers in routine tasks.

On average across OECD countries, occupations at highest risk of automation account for about 28% of jobs. Workers in occupations with the highest shares of automatable skills and abilities continue to be low-skilled, young, and male.

An increasing share of workers in OECD countries are legally employed by one firm but in practice work for another. Cleaners, security guards and cafeteria staff are examples of occupations where workers often physically work on the premises of one firm, but their legal employer is a third-party support services firm. Such third-party employment relationships are often referred to as “domestic outsourcing”.

Employment in the administrative and support services industry provides a broad measure across countries of the growth of domestic outsourcing. This industry captures general business support services provided to other firms. Examples include temporary employment agencies, travel agencies, as well as cleaning and security services to firms. Over the past 20 years, employment has grown briskly in the administrative and support services industry. From 1995 to 2005, the share of total employment in this industry grew from 3.6% to 4.9%. From 2005 to 2018, it continued to rise to 6.3% of total employment.