Artificial selection

Are HR artificial intelligence platforms discriminating?

Companies are using artificial intelligence (AI) platforms in the interview process to simplify and speed up hiring. However, the actual use of AI highlights concerns that discrimination caused by flaws in technology and associated processes results in suitable candidates missing the shortlist.

The coronavirus pandemic has really had an impact on the global job market, with almost all verticals reducing recruitment in their 2020 recruitment. In particular, looking at the UK graduate market, the most affected segments have been finance, B2B companies, media and retail. For example, to quantify this, the number of graduates recruited in 2020 was 12.3% lower than the previous year, with major employers in the country recruiting 3,700 fewer graduates compared to the previous year.

However, this is only the graduate segment, and if we consider non-graduate positions in verticals like retail, the implications are far greater. At the same time, there are also more candidates applying for each job offer. As an example, a vacancy for an investment banking analyst in the UK would easily attract over two hundred applications, with at least 90% of applicants having acceptable higher level qualifications and undoubtedly the right references. .

So, the killer question: how do HR recruiting functions make the selection process as fair, efficient and fast as possible – and can artificial intelligence help?

Has recruitment ever been fair?

Increasing emphasis on environmental, social and corporate governance (ESG) compliance measures by businesses (including corporations) globally, emphasizing diversity and equality , you would have thought that today’s recruitment process was fair for everyone, regardless of race, gender and origin.

However, independent studies such as that conducted by the Institute for Social and Economic Research at the University of Essex in the UK have found that, for example, UK graduates from ethnic minorities were between 5% and 15% less likely to be employed than their white British counterparts. peers six months after graduation. This is just one research study and there are many more. A key contributing factor is that it’s not so much about having the right processes and the best ESG-compliant metrics in place. But it is this unconscious bias in the interview process that is a contributing factor.

Surely artificial intelligence in recruitment will solve all this?

AI in recruitment, especially interview screening, aims to deliver many benefits on paper. First, meet the KPIs within the company, in terms of time taken and securing a good hire. Again, the focus is on finding the right people who can help make a business successful. Key attributes that AI platforms and associated value-added software can provide include managing the entire initial application process, screening all CVs and application documents. It also provides the automation of interviews that would normally be carried out by recruitment staff, through the use of predefined questions or video chats, personality profiling and ultimately scoring, ranking and shortlisting. candidates. Some AI platforms in the market also provide analysis of facial movements, accents, word choices and appearance of potential candidates, before ranking them according to their suitability.

Some of the key components of AI as a whole may include learning, reasoning, problem solving, perception, and language understanding. In addition to this, AI interview platforms work by comparing interview candidates and responses given to data sets, and through algorithm score assessments, sometimes analyzing how humans evaluate candidates.

However, with all the good that artificial intelligence brings to society and businesses, it is ultimately based on learning models mimicking a human, and in applications like recruitment where unconscious bias in the interview process come into play, technology has many flaws, in addition to technological and moral.

The UK’s BBC broadcast channel recently aired a show in which AI was apparently used not in the hiring process, but in the firing of Estee Lauder at its MAC Cosmetics business, and a responsible case is currently pending involving the former employees. Other real cases at issue include how race, gender, emotion and regional accents were misinterpreted by facial analysis technology in the recruitment process.

With these drawbacks of artificial intelligence in recruitment and the pressures on society in the labor market, employers and especially large companies looking to streamline processes need a new course of action to make fair recruitment for all, and to do that they may need to go back to basics and address “very real” issues in society and business about discrimination based on unconscious bias. Only then will technologies like AI in recruitment have the impact its vendors claim to have.