Does unconscious bias training really work?

Does Silicon Valley’s latest quick-fix for its diversity problem actually work? Or does it just reinforce biases?

I've read some good pieces lately on the state of diversity -- or lack thereof -- in Silicon Valley. This long-form Atlantic piece, "Why is Silicon Valley so Awful to Women?" does a solid job of laying out the challenges to a fully inclusive tech industry, as does this two-part PBS NewsHour story, "How Silicon Valley is Trying to Fix its Diversity Problem," based on similar reporting and with associated video.

One thing stood out to me, though, as I was reading the Atlantic piece: Unconscious bias training might not work. In fact, it could actually make people more inclined to bias, since it normalizes bad behavior through its message of "everyone holds biases, and there's nothing wrong with that."

And then, this quote from Mike Eynon, a tech executive who wrote that unconscious-bias training simply serves to "'make us white guys feel better' and lets the '"privileged realize everyone has bias and they aren't at fault," while nothing changes for discriminated groups.'

Well, damn! So, what is working, if anything? Well, the article goes on to explain that a combination of structural changes and culture shifts, along with an "anti-bias checklist" of approaches and technology (like Textio, which gender-neutralizes language, or Blendoor, which completely anonymizes resumes to remove gender, race and other bias-triggering factors) has the best shot of making a difference; there's no single silver bullet.

But it really all comes down to culture, in the end. "Diversity consultants and advocacy groups say they are frustrated by companies' unwillingness to change core parts of their culture," the Atlantic article says, like the infamous practice of whiteboarding during a technical screening, which is proven to unfairly disadvantage women and people of color in interviews. Whiteboard interviews contribute to "imposter syndrome," they discriminate against those who don't have time to cram obscure algorithms in their heads, and they don't actually test many of the skills you'd need to be successful in a programming role.

A few weeks ago, a Twitter thread started by Ruby on Rails creator David Heinemeier Hansson, invited programmers to publicly "confess their sins" in an attempt to demonstrate that good, successful coders don't necessarily excel at these kinds of tests. But the thread also served to highlight just how sexist the industry still is when women started to chime in:

"Hello, my name is Erica. If I told you what things I still have to Google, you'd try to use it as 'evidence' that I'm not a 'real' engineer," one woman Tweeted.

Burn.

And so, it would seem we're back to square one. No matter what tools, tricks and tech you throw at this problem, if you don't fix the underlying, systemic problems of sexism, racism and homophobia, what you have is a whole lot of high-tech Band-aids.

Related Video

SUBSCRIBE! Get the best of CIO delivered to your email inbox.