Guest Blog: Taking Responsibility for AI’s Gender Bias

Guest blog by Arran Stewart, Founder and CVO of Job.com

Last year Amazon, arguably the largest and most successful tech company in the world, publicly scrapped the artificial intelligence hiring tool it was planning to use for company wide recruitment. The tool, which the company had been working on since 2014, was found to have an unconscious bias against women, not selecting their resumes for consideration for “male dominated roles” despite being equally or more qualified than male counterparts.

Now this isn’t the first time in the last decade or so that AI-backed recruitment and mainstream discussions about equality and fairness have intersected. Still, it’s almost like the recruitment industry would rather ignore the gender-biased elephant in the room. I’m certainly not the first and won’t be the last to bring up this topic, but as an executive in recruitment, especially one with the jackpot of my privilege, I feel it’s my duty to do what I can to acknowledge and address the issue as best I can.

Gender bias in science, technology, engineering and math

It’s undeniable at this point that STEM jobs have a gender parity problem and seem to have always had. On paper, women make up about 50% of the STEM workforce, however this is largely due to their overrepresentation in healthcare. In the tech fields we’re discussing, those that build the AI that shapes our world, women are woefully underrepresented. Only 14% and 25% of the engineering and computer fields have female workers. To cherry pick some big names, only 27% of Amazon’s global leadership is female, while 80% of Microsoft’s leadership and technical staff is male.

Programmers, software engineers and designers build AI based on how they see the world. Despite our best efforts to be egalitarian and objective, it’s still impossible to completely remove the “human” element from artificial intelligence. People are inherently biased therefore the technology created by them, which is supposed to mimic the way that people think, will never be truly objective. However, that doesn’t mean that companies are blameless in – if not actively creating – passively allowing an environment and culture to prevail STEM that makes women reluctant to join the industries.

Doing the work

As a technologist, I’m committed to using tech to make the world a better place. As someone with a platform, I think it’s critical for myself and others like me to lead the call for diversity within our industry. We’ve kept our heads to the keyboard in the last few decades, in an endless arms race with each other to build the next big thing; so much so that we haven’t stopped to examine what our technology has done, especially to the world it’s increasingly shaping. We know already, from a technical standpoint, the fixes we need to make; the codes we need to write, systems we need to create to create a better artificial intelligence: we just need to do the work now.

By 2020, the majority of major global companies will use AI for customer and business analytics. This real world data informs the algorithms that teach AI and inform product direction, marketing strategies, resource allotment and every other aspect of business. To make sure that the analysis of this data doesn’t just continue to perpetuate the same biases of yesteryear, we need to build awareness into AI. We need to put more diverse voices behind the scenes so that the technology can have a broader perspective. This not only means inclusivity across gender and racial lines, it means taking a holistic approach to technology pulling in expertise from an array of fields and not just STEM.

Education

We need to empower and educate women and other minorities in our field, heed their voices when they speak up about their experiences and work to make sure their issues are addressed. We need to put them in leadership positions and push for fair representation across all sectors. As people on top, we must hold doors open for others. It’s not a company-specific issue, these biases affect us all; limiting our potential and keeping entire populations from seeing themselves reflected in our world.

As the industry that most closely influences who works where, it’s critical that we make sure our tech gives everyone a fair chance. Male executives – myself included – must be ready to do the work to make their companies – and the tech they’re creating – diverse and inclusive. I would love to start a US council driven by leaders in our industry aimed at creating an industry standard of artificial intelligence testing. No matter how competitive the recruitment landscape may be, as technologists who are driven to serve and improve hiring, we have an obligation to be sure we strive for equality.

About the Author

Arran Stewart is a global keynote speaker and the co-founder of blockchain recruitment platform Job.com. Relying on a decade worth of experience in the recruitment industry, Arran has consistently sought to bring recruitment to the cutting edge of technology.