Arnd Visarius

Job search discrimination – does AI hold the answer?

Arnd Visarius
Arnd Visarius
published 01/10/2020

During one of our jobvalley Campus Expert Talks, we touched on the topic of how artificial intelligence (AI) can help to effectively combat discrimination when hunting for a job. Carsten Dickmann was joined by Larissa Fuchs from FAIR.

Nowadays, artificial intelligence can be found in every aspect of life. It’s hidden in social media algorithms that decide which content to expose us to and found more visibly in chatbots used for customer support or with facial recognition used by mobile devices on a daily basis. These applications all have one thing in common: boosting efficiency and improving customer experience. Algorithms may also have the power to put an end to discrimination.

Discrimination in the recruitment process – how things currently stand

Job search discrimination remains a major issue. According to a survey commissioned by the Federal Anti-Discrimination Agency, 23.9 percent of the 4,000 reported cases of discrimination were related to the recruitment process. A total of 41 percent of the counselling cases handled by the Federal Anti-Discrimination Agency are associated to working life, with one third related to access to work. Prejudices and stereotypes are commonplace in human relationships; searching for a job is no exception. Last but not least, we need to vastly improve our approach towards unconscious biases or gender biases. 

Can AI help reduce job search discrimination? 

Artificial intelligence is still in its infancy in many areas; it requires further development in the field of recruiting. Nevertheless, we can already plausibly say that AI may have the ability to become a key element in hiring staff with ongoing advancement. The 2020 pandemic has helped digitisation come on leaps and bounds in numerous sectors and changed the way we work. The same can be said for recruitment. AI now has the potential to review candidate profiles for behavioural attributes and personal values along with the applicant's compatibility with the hiring company – their cultural fit.

jobvalley Campus: Larissa Fuchs in conversation

The FAIR framework (Fair Artificial Intelligence Recruiting) offers an interesting approach to the use of AI to prevent discrimination and pave the way to more equal opportunities in suitability diagnostics. During a jobvalley Campus Expert Talk, Carsten Dickmann, Team Manager Marketing at jobvalley, explored to topic with his guest Larissa Fuchs. Larissa Fuchs is a PhD student at the University of Cologne who specialises in researching labour market discrimination.

What is FAIR? 

FAIR is a research project headed by candidate select (CASE) and the University of Cologne conducted between 1 January 2020 and 31 December 2021. FAIR aims to both systematically detect and eliminate factors used to discriminate in recruitment through the use of algorithms. The project is funded by the EU and the state of North Rhein-Westphalia. jobvalley is an associated partner of FAIR. 

What does FAIR hope to achieve with its AI? 

FAIR's mission is to develop and promote a demonstrably fair algorithm to improve how things currently stand in terms of discrimination. The FAIR algorithm helps avoid the application of unconscious biases or gender biases in the screening phase. It offers an AI-based recruitment tool that pairs efficiency with anti-discrimination.

Is AI 100% objective?

In theory, artificial intelligence is free from prejudice, making it more objective than us humans. However, algorithms – including ones used to screen candidates – rely on the input or information we give them. Accordingly, they are trained on the basis of existing data, but need to nevertheless exclude existing biases. FAIR aims to counteract this with the FAIR Index, which is capable of detecting both human-made and algorithm-based discrimination by demonstrating which people were ‘unfairly’ hired and whether any suitable candidates were not taken into consideration. Deviations may not always be problematic in nature. However, if certain demographics are found to be consistently discriminated against, this may indicate systematic discrimination. The FAIR Index can identify algorithms that have adopted this kind of discrimination. In response, they can then be modified to increase the fairness of the recruiting process. 

jobvalley summary

Recruiters may be professionals, but they are still human. Despite their experience and professional training, they will never be able to completely rid themselves of (unconscious) biases. Indications of your ethnic origin, gender or even the photo in your application may influence their decision; it's not without cause that these elements are specifically banned from applications in certain countries. AI can help recruiters screen CVs and prevent (potentially better) matching profiles from being excluded early on in the process, ultimately giving more applicants a fair chance of being hired.

At the same time, these kinds of system solutions have the potential to help companies become more efficient and flexible. Two major pluses in the day and age of remote work. We also rely on the benefits of modern technology combined with working with specialists in their fields, as demonstrated with the new jobvalley customer portal.

Arnd Visarius
About the author

Arnd Visarius

Since I've always been passionate about writing, it didn't take much effort to jump out of my studies into the fresh water of online editing when the opportunity presented itself. After working in Dortmund, Witten, Cologne (other side of the Rhine) and North Miami Beach, I'm now aboard as Content Manager at jobvalley.