12 Jun 2017 16:49 IST

Weeding out biases with tech

Pic credit: Aleutie/Shutterstock

When machines take over appraisals and recruiting, will we see a workplace without inequities?

If you are human, you are bound to have biases — conscious or unconscious, positive or negative. However much you try to be fair, these snap judgements are reflected in your hiring decisions or in annual performance appraisals. Let’s face it: better looking candidates or those with a force of personality always tend to score more in interviews.

But what if machines did the candidate screening? Or machine learning software checked for unfair or inequitable compensation?

Towards equality

The first wave of using artificial intelligence to do resume matching, screening and shortlisting candidates has already begun. At its annual SAPPHIRE NOW conference held in May, global enterprise application firm SAP announced that it would use text mining and machine learning to help companies review job descriptions, performance reviews and similar people processes to search for potential bias and suggest changes to encourage equity.

Through its acquisition of Success Factors, an HR software firm, SAP already has the tools in place to check gender bias. Success Factors’ software looks at language used in a job description and recommends changes if it feels a gender bias is present.

Discouraging words

According to studies, many women hold back from applying to jobs simply because of the inherent bias in the postings. For instance, descriptors such as ‘rockstar’ or ‘ninja’ are words that deter female candidates from applying. Success Factors’ software suggests other nouns.

Similar red flags are raised during performance review modules. What this does, explains a SAP India spokesperson, is it empowers people to create descriptions that contain gender neutral words and over time, help them weed out the habit.

The software also does data crunching to match resumes to jobs. And unlike humans, who may get swayed by the personality of the candidate, the software focuses on the needs of the jobs and looks at the candidates’ skills.

SAP has been using these tools in-house to improve gender and ethnic diversity, and to remove inequities, before providing it to clients. To close a pay gap that was highlighted through these tools, SAP gave quite a number of its employees’ raises.

Host of companies

SAP is not the only one. A host of companies are now providing the tech to eliminate bias or to broaden the candidate search pool and improve diversity. For instance, recruiting platforms such as Power to Fly and Entelo help companies find a wider and more diverse pool of candidates by trawling through information on social networks. Entelo has helped its client Schneider Electric access more than 300 candidate profiles, particularly technical talent.

Then there are platforms like Textio that use augmented writing to create effective job listings. Textio analyses hiring outcomes of millions of job posts and then rates the performance of a company’s listing, helping it improve.

Weeding out discrimination

Some of the new software coming out of the labs of HR tech start-ups even gauge sentiment. Take San Francisco-based Kanjoya (now acquired by The Ultimate Software Group), which uses natural language processing to spot discrimination. It can ferret out what people are really thinking when they fill employee surveys.

For instance, if leadership is mentioned in a negative light in the survey forms of younger employees, the software can throw up the fact that the company is not treating its freshers as well as it treats middle rung or senior employees.

The good news is that a whole lot of HR tech is now coming up to remove inequities at the workplace. But that does not necessarily mean that something like recruiting or performance management, which is all about people-to-people connect, will be taken over by machines. It will simply empower managers to be more mindful.