The many uses of artificial intelligence for human resources must be regulated

As the areas in which artificial intelligence (AI) is being used are becoming more varied and numerous, it is already being applied to HR as well as to the public sector. The latter is a major employer and has been won over by the many opportunities for procedure automation and analytical information processing.

With Germany at the helm, the OECD rolled out a programme to measure and analyse the impact of AI on work, innovation, productivity and skills in 2020 (1). The aim is to help governments manage the transition by ensuring responsible and human-centred use to improve individual and societal well-being.

In Belgium, the federal services are conducting a survey of their staff to assess the extent to which AI is being used and to identify the remaining questions or reluctances. Drawing on the results, strategic recommendations will be made in the near future.

In Austria, job seeking can now be optimised. The civil service job exchange now has a CV parsing functionality to make these searches easier and personalised. Uploaded CVs are analysed by a system that codifies and structures the information so as to show offers that match these criteria. This means that the selection is more finely-tuned and more accurate.  

The first “virtual job interviewer” has been introduced in a Swedish municipality as a direct result of the COVID-19 pandemic. In Upplands-Bro, the main feature of the Tengai robot, which was developed by Stockholm's KTH Royal Institute of Technology and a startup, is conducting interviews using data that allegedly mitigates unconscious bias by removing the human influence.

It is highly likely that civil service HR information systems will make use of this technology and South Korea is already a trailblazer in this respect. E-Saram is so much more than a management application as it helps with HR policy implementation by using a number of decision-making assistance modules. HR advisers carry out simulations (defining priority skillsets for positions subject to hiring difficulties, modelling career paths to help guide and select applicants or arranging for knowledge transfers by identifying possible partners).

Use of AI technologies will need to be rapidly incorporated into training given to civil servants and falls outside the confines of digital skills as it has a strong ethical content. In the United States, federal employees will receive AI certification after taking a special course and undertaking to strictly comply with a code of ethics.   

A large number of projects – some of which may be controversial – are already being initiated. For instance, in Spain, the autonomous region of Andalusia is looking to introduce a “civil servant genome”. Its designers explain that this would involve collecting comments posted on social media to “better understand interests and provide access to positions to which it can adapt”. The project is facing opposition from professional bodies.

Taking advantage of the many opportunities offered by AI is therefore a trend that is gradually gathering ground. In the United Kingdom, the regulatory authority responsible for upholding information rights has warned the civil service about the risks, or even abuses, which could be caused by the mainstreaming of intensive automation. Algorithms result from human programming and, as such, do not guarantee either equal treatment or non-discrimination.

¹ To date, 46 countries have approved the Recommendation of the Council on Artificial Intelligence which was adopted on 22 May 2019.