How online recruiting can create bias

Online recruitment software are a godsend to HR managers everywhere, as they streamline the process of hiring the ideal candidate. However, they may not eliminate bias completely.

Online recruitment is touted to be an objective, fast, efficient, time- and money-saving technology, which can ease the process of recruitment for both the employers and the candidates. In addition, it can give employers a chance to recruit the best talent from a wider pool. Nowadays, even artificial intelligence (AI) is quite commonly used across recruitment software, using data from much deeper micro analyses.

Online recruitment may replace offline altogether, but there is still a long way to go. Technology has not evolved as much as it should to eliminate biased decision making.

To correct the machine bias, it is necessary to correct the human factor first. We all have unconscious biases within, which we may or may not be aware of. These can manifest in the decisions we make. Recruitment software require a lot of parameters, which are to be defined by the employers, and in doing so, some unconscious bias is bound to creep in. For instance, if a company has been hiring from a particular set of colleges and experiencing good results, it may input those colleges as the ideal requirement. The software will then award candidates from other colleges a lower ranking and possibly miss out on better talent. This is called the clustering effect, which means believing all good things come from the same cluster.

Again, the hiring medium itself may lead to biased opinions of the candidates. For instance, if, during a video interview or a telephonic one, the connection is not pitch perfect, then the candidate’s voice and tone may get distorted. This is quite a realistic scenario. And if the job requires excellent communication skills or a certain manner of speaking, then it may reflect negatively on the candidate.

“I would use the word ‘preference’ and not ‘bias’. Moreover, it is true that online recruitment technology was inefficient before, but that is not the case anymore. Nowadays, it has really evolved and there is better predictability because the human element has been removed. There is now more data to refer to, which has brought down the rate of preference hiring,”  asserts praveer priyadarshi, fomer CPO, jindal stainless.

Praveer Priyadarshi, former CPO, Jindal Stainless, disagrees. According to him, the kind of technology we have today and the rate at which it has advanced should be appreciated. “I would use the word ‘preference’ and not ‘bias’. Moreover, it is true that online recruitment technology was inefficient before, but that is not the case anymore. Nowadays, it has really evolved and there is better predictability because the human element has been removed. There is now more data to refer to, which has brought down the rate of preference hiring,” he asserts.

Recruitment technology uses data from all previous hires, and companies provide the same in huge volumes. It acts as a good reference point to select the ideal candidate type from amongst a group of possible hires. However, the data provided can again be a source of bias. The software is trained in the data provided, which may contain biased human decisions or the second-hand effects of those decisions. Even if the opposite is possible, it does not rule out the possibility of biases occurring.

Amit Das, CHRO, Bennett Coleman, concurs and explains that the concern for bias depends on the kind of role or functionality we are looking at. When it comes to volume recruitment and the role is a functional one, then despite the possibility of bias, there are no qualms in using the online tools. “It is only when the role requires complex skills and the ‘how’ part of the performance matters, that there is focus on offline recruiting,” he elaborates.

“It is only when the role requires complex skills and the ‘how’ part of the performance matters, that there is focus on offline recruiting,” elaborates Amit Das, CHRO, Benett Coleman.

Mass recruitment can take a heavy toll on the hiring managers if done offline, and it makes sense to conduct it online. Then again, since mass online recruitment can let some bias seep in, it may deprive the already underprivileged class of people in the society from getting a job, no matter how functional.

There are unskilled or less skilled minority groups in the society, such as the PwDs or the LGBTQI community, who have been struggling to earn a regular income. If the system rejects them, they are excluded from most jobs. In such cases, despite commitment to diversity and inclusiveness, an organisation may unintentionally end up excluding these communities.

“The purpose of going online is to lessen the burden on the hiring managers, and also reduce the cumbersome job of filtering. But the question we should ask is whether those filters themselves are objective or not,” says Sunitha Lal, CHRO, Ather Energy.

Sunitha Lal, CHRO, Ather Energy, agrees that when defining parameters to create the ideal candidate type, it is easy to get carried away. “The purpose of going online is to lessen the burden on the hiring managers, and also reduce the cumbersome job of filtering. But the question we should ask is whether those filters themselves are objective or not,” she concludes.