Amongst the many ‘plug-in’ solutions that organizations can use right now to deliver business efficiencies, AI recruitment tools are perhaps the most obvious. There are numerous systems on offer that will take your job requirement and use machine learning to find relevant candidates. What’s not to like?
AI sourcing is in theory a great way to speed up and ease the recruitment process—particularly helpful during the pandemic when job applications are likely to soar. However, machine learning involves the use of algorithms that search for patterns based on past data. Recruiting for the future your organization may need to look for new types of talent and the past may be a poor guide.
In this article, first published on St Gallen University’s Vista platform, Prof. Dr. Gudrun Sander weighs up the advantages and potential dangers of using AI recruitment tools:
The traditional recruitment process is time-consuming and can take months. Applications must first be screened and evaluated. Then, in-person interviews follow. Various online tools show that this can be done faster. HR managers enter the profile requirement into matching platforms and these find suitable candidates within a very short time. The selected applicants receive automatic invitations, e.g. for an initial video interview. Many companies already take advantage of online application tools. Artificial intelligence, however, has not been commonly used so far. Upon the online applications, programmes can scan the data and attach a weight to e.g. education or career. The algorithm then creates a shortlist and suggests applicants to recruiters. The recruitment time but also the personnel effort can be reduced massively. Besides, with free time at hand, HR managers can focus on other aspects, e.g. on the final interviews.
But not only the time and personnel savings are in favor of artificial intelligence, but also the “neutral” nature of computers, i.e. they are free of emotion and prejudice. But here is also the catch. All programmes will have to be programmed by humans in the near future, who in turn are not free of prejudice. On top of that, every artificial intelligence needs a data basis from which it can learn and derive decisions. Existing employees or application files serve as a basis. If in the past, mainly men were hired for senior management positions, the system learns this as a positive selection criterion. Hence, it will increasingly propose men for management positions. The machine is trained and, depending on the programming, the algorithm “worsens” towards reinforced stereotypes. In the medium term, this would then lead to an even more pronounced occupational segregation of the labor market and de facto increased inequality of opportunity.
A further disadvantage of artificial intelligence in the recruitment process that is not obvious at first glance is that only persons who can be found online with their profile are suggested. In addition, outdated terminology, e.g. in connection with education that dates far back or a job title, is often not recognized. In this, re-entrants or people with non-linear careers are often not considered, although their skills might fit a position.
So how do you find the right talents for your company? HR managers must take a close look at the individual steps in the recruiting process and consider from a diversity perspective at which process steps the use of artificial intelligence makes sense and where the experience of recruiting experts is superior. If you want a more diverse workforce, you can, for example, adapt the tools to take the criterion “diversity” into account. You could also expand the criterion “education” (e.g. other technical education instead of just mechanical engineering) or years of experience. How do such adjustments affect the applications? You can gain experience and the recruitment process will be of better quality. It remains important that HR and line managers make the final decision but are supported in the process by artificial intelligence to increase efficiency.
Therefore: A combination of human and machine will be important if recruiting processes are to contribute to more diversity and inclusion in companies.