Could AI end the workplace diversity struggle?

Listening to Tom Loeffert of SAP speak yesterday at TechXLR8 — the last day of the week-long London expo —  I was struck by something strangely unfamiliar: an outlook on AI that doesn’t centre around speeding processes or creating a new digital workforce. Or ending the world. No, instead Loeffert spoke about AI’s ability to drive gender diversity in the workplace, an altogether unexpected outcome from a typically more ‘practical-minded’ concept.

Over the course of a short and sweet 20 minute presentation we learned about algorithms and data processes, used by SAP, that counter employer’s unconscious bias. The example Leoffert used goes like this: if 40% of the people who apply for a job in tech are women, a member of that group has only a paltry 5% chance of getting hired. This isn’t right, but it comes down to an unconscious bias on behalf of the employer – SAP have algorithms that, via a process of deep learning, can actually correct this bias and ensure a proper ratio is employed.

Put plainly: AI has the potential to counteract a person’s natural tendency to discriminate. An endearing and unprecedented application for the robots, even if it sheds a somewhat unflattering light on us.

SAP seem to practice what they preach. Clearly they’ve been using their own algorithms as last year they became the first multinational technology company to be awarded the worldwide Economic Dividends for Gender Equality (EDGE) certificate. Something that may, or rather should, sound redundant in 2017, but in the tech industry is groundbreaking. Indeed, the world of tech is almost sickeningly male dominated, with last year’s ‘The Industry Gender Gap’ report highlighting a mere 26% of women worldwide holding jobs in the industry. In the UK, it’s 17%. I for one was not expecting AI to be providing the best and seemingly only answer for this.

Gender is just one facet of the diversity issue, however – and unlocking genuine employee diversity and its massively beneficial business outcomes means going beyond nominal, quota-filling approaches. After all, it’s perfectly possible to employ a workforce comprising a wide range of  different nationalities and yet still have an office full of straight, white males. For AI to help us construct a truly diverse workforce, it needs to be developed in a nuanced way that addresses more than just unconscious bias.

For now, SAP’s approach offers a tantalising glimpse of the big part could AI play in combatting the overly one-dimensional nature of the global workforce at large. It’s an exciting idea, especially since it gives us a refreshing perspective on the applications of Artificial Intelligence. Supposing it isn’t merely a pipe dream, we may have the answer to our employee diversity problems sooner than we think.