Lessons from Davos for building a diverse team

LinkedIn's co-founder Allen Blue tackles the issue of unconscious bias in AI.

by Allen Blue
Last Updated: 01 Feb 2019

Last week, at the World Economic Forum, the rise of Artificial Intelligence was one of the most exciting topics up for discussion. I was there and joined a number of panels, speaking up on a issue that I think very important - the value of AI to human society, and the importance that we do it right, and why that makes it urgent that we respond to the gender imbalance in the workforce when creating AI.

Prior to the Forum, we worked with WEF on their Global Gender Gap Report, providing insights from our platform to understand the problem. We call this tool the Economic Graph - a digital map of the global economy - which draws upon the experiences of nearly 600m LinkedIn members worldwide to identify how the economy is changing, or not changing.

And that’s the issue in the field of AI. According to LinkedIn insights, 78% of the workforce with AI skills are men, and as women are gaining AI skills at the same rate as men, that gap isn’t closing. Even with the best intentions, unconscious biases exist. They exist in the engineers building AI tools, they exist in the data sets that they use to build models, and they exist in their judgement of whether or not the prototype AI returns a ‘correct’ response.

While raising awareness of such biases is important, it only goes so far - the real solution is to increase the diversity across the team building the AI. Prior to traveling to Davos this year I met with some of the women in our own engineering team and the solutions they advocate have applications beyond the AI industry - they offer value to anyone addressing a diversity problem, from heavy engineering, to the aviation industry, to sales.

Cycles can be vicious, or virtuous

We’re in a vicious cycle - when a profession has too few women, there are fewer examples of success, there are fewer women in leadership, there are fewer women doing the hiring - all leading to more unconscious bias. The result is an issue that many industries are struggling with. The key is to disrupt the cycle - but how? 

Throughout my week in Davos, the lack of Women in STEM (the so called ‘pipeline problem’) has been called out as the largest issue. It raises the question where does the female talent go, and why. But one of our own engineers was keen to blast that idea open: it’s not true of all sciences: the life sciences disciplines are majority women, for example.

Her point of view was that at an early age girls and boys - and I’m not the expert on the nature vs nurture debate in this - look for different things. Boys look for competition and individual success, girls respond more to cooperation and contribution. The life sciences, she argued, attract more interest from the girls that grow up because they offer that combination of doing something good, for the group.

The problem is the stereotypical image of an AI engineer - a young man working alone in a messy dorm room - doesn’t suggest that sense of cooperation, despite the fact that real engineering teams, working together on a common goal, can create a tremendous atmosphere of community. We all need to do a better job showing young women that there are many ways to contribute to a greater good.

We need to be open to people with adjacent skill sets

While we can hope to inspire the next generation of young women, there are things we can do sooner - at LinkedIn, to diversify our teams working in AI, we’ve begun to unlock talent from adjacent skills areas, finding people in fields with a stronger gender balance who have the mindset, intelligence and skill, but perhaps not the PhD in Computer Science. Our AI teams have found fantastic people in areas of biology, where the core of the task - understanding patterns - is common to both. Beyond AI, we’re increasingly seeing women move from areas of the business such as HR and marketing into sales.

We’ve got more work to do, but I left Davos with a confidence that there are things we can do today, and a willingness to do them across technology and other industries to ensure that opportunities such as this are open to all.

Allen Blue is co-founder and VP product management at LinkedIn

Image credit: kalhh/Pixabay

Tags:

Allen Blue recommends

Will AI make us redundant?

Read more

Find this article useful?

Get more great articles like this in your inbox every lunchtime