Chatbot or human – who makes the better recruitment choices?

Jessica Ochmann and Prof. Dr. Sven Laumer
Prof. Dr. Sven Laumer and his team believe that chatbots are dialogue systems that are suited for making processes in the area of human resources more efficient, but raise their concerns that chatbots need to become more intelligent and capable of responding to users in a more individual manner. (Image: FAU/Georg Pöhlein)

Team of researchers at FAU has published study on AI in human resources

One available position, 1000 applications. More and more large companies are using artificial intelligence (AI) to increase efficiency in human resources, in particular when looking to recruit suitable staff. One area where chatbots can prove useful is to give prospective candidates an initial impression of what the job will entail and to help them assess how well suited they actually are for the position. The Adecco Stiftung für Arbeit und soziales Leben (Adecco Foundation for Labour and Social Affairs) has financed a three year project at FAU on the topic of ‘Artificial intelligence, chatbots and recruitment’. It investigated the relationship between humans and machines in human resources, gaining valuable insights into different topics such as the ethical aspects of these digital systems and the extent to which people are willing to accept them.

Researchers at the Endowed Chair of Digitalization in Business and Society at FAU carried out interviews with potential users and stakeholders in the area of human resources as well as with experts in the field of AI. The team of researchers from Nuremberg led by Prof. Dr. Sven Laumer have now summarised their results in a report that is of both practical and academic relevance. The researchers believe that chatbots are dialogue systems that are suited for making processes in the area of human resources more efficient, but raise their concerns that chatbots need to become more intelligent and capable of responding to users in a more individual manner. Even though algorithms can help human resources departments come to a decision, recruiters still tend to be sceptical about relying entirely on data. The study indicated that users prefer humans over machines when it comes to sensitive matters concerning personal data or when making recommendations that are of relevance to the future. Another interesting finding was that AI is more widely accepted if decisions are seen as being fair. Bearing this in mind, the researchers developed and tested a multi-dimensional fairness model. The study also suggested that applicants tend to regard companies that offer digital recruitment as an innovative and attractive potential employer.

However, one aspect which must still be addressed is the active avoidance of discrimination. It is often the case that discriminatory features (such as demographic characteristics) are applied in models that are used as the basis for generating recommendations. This underlines the importance of a high degree of transparency, which would in turn increase the accountability of AI-assisted recruitment.

The project was funded by the Düsseldorf-based Adecco Foundation, which focuses on new approaches for work and social affairs. According to the director of the Adecco Foundation, Janine Bischoff, ‘the future of work has been one of our key issues from the outset.’ ‘We were happy to provide funding for this exciting topic.’ In order to be able to continue research at FAU into digitalisation and follow on from the research started by the project funded by Adecco, the Dr. Theo and Friedl Schöller Stiftung will provide funding for a research laboratory focusing on the future of work from February 2022.

Further information

Prof. Dr. Sven Laumer | Jessica Ochmann, M.Sc.
Schöller Endowed Chair of Digitalization in Business and Society
Institute of Information Systems
wiso-wi-dwg@fau.de