Skip to content
Business

We can’t “de-bias” our brains but we can change our systems

By designing smart systems, we can help ourselves live up to our best intentions — and perform even better in our workplaces.
An open magazine reveals a woman wearing a hijab and headphones on one page, with de-bias technology illustrations mapping the faces of three individuals on the opposite page.
Adobe Stock / Prostock Studio / Chaytee
Key Takeaways
  • When constantly reinforced, our biases can turn into a self-fulfilling prophecy.
  • Unconscious bias at the individual level is extremely difficult to change.
  • However it is possible to de-bias organizational processes, environments, and policies.
Sign up for the Big Think Business newsletter
Learn from the world’s biggest business thinkers
Excerpted from MAKE WORK FAIR: Data-Driven Design for Real Results By Iris Bohnet and Siri Chilazi Copyright © 2025 by Iris Bohnet and Siri Chilazi. Reprinted by permission of HarperCollins Publishers.

Our biases have far-reaching consequences for others, and when constantly reinforced, they can turn into a self-fulfilling prophecy. By age six, research suggests, American children have absorbed the association of intellectual brilliance and genius with men more than women. Heartbreakingly, six-year-old girls are already less likely than boys to believe that members of their gender are “really, really smart.” These stereotypes have a direct influence on children’s behavior: girls are less likely than boys to express interest in novel games that are intended for children “who are really, really smart.”

In a similar vein, economist Michela Carlana showed that in Italian middle schools, girls underperformed in math when they were assigned a math teacher with strong unconscious beliefs that boys are better at math than girls (as measured by the most widely used tool, the Implicit Association Test). Furthermore, these girls ended up self-selecting into less demanding high schools, which influenced their career options going forward. The gender stereotypes of the teachers induced the girls—especially those whose math skills were lower to begin with—to doubt themselves, leading them to fail to reach their full potential. One can only wonder how much talent we collectively are missing in the world when we start to lose out on it at such a young age.

Book cover with a purple background. Text reads: "Data-Driven Design for Real Results. Make Work Fair. Iris Bohnet & Siri Chilazi.

The pernicious effects of bias also play out in the workplace—with negative consequences for individuals and organizations. In grocery stores, female managers were found to allocate disproportionately more time to tasks that allowed them to disprove negative gender stereotypes about their lack of commitment or competence. This meant spending more time on the grocery store floor on tasks they could do visibly in front of subordinates. Unfortunately, it resulted in the female managers spending less time on more invisible, yet equally important, office tasks, which in turn hurt their and their departments’ performance. In medicine, where female surgeons were not afforded the same respect and authority as their male counterparts due to gender stereotypes, they behaved more deferentially toward the female nurses they worked with, such as by helping the nurses with their tasks and befriending them outside the operating room. This “managing down” hurt female surgeons’ performance.

People’s actions change when the systems surrounding them change—even if their unconscious or conscious biases do not.

If you are now thinking that eliminating bias from human brains is necessary to create fairer workplaces, you are not alone. Unfortunately, that approach is hardly feasible because unconscious bias at the individual level is extremely difficult to change. At best, some programmatic interventions have managed to temporarily affect people’s attitudes and/or their awareness of bias. But there is little indication that any change in unconscious bias awareness has led to sustainable changes in behavior, according to the most comprehensive meta-analyses synthesizing evidence from hundreds of studies over the last two decades. Simply put, de-biasing humans is incredibly hard, and might well be impossible.

Try Big Think+ for your business
Engaging content on the skills that matter, taught by world-class experts.

What we can do is de-bias the organizational processes, environments, and policies that we have created and that contain built-in bias. Consider the commercial facial recognition software that can accurately identify the gender of a white man in a photo, getting it wrong only 1 percent of the time, but that fails to recognize darker- skinned women up to 35 percent of the time. This software was built and tested by mostly white, mostly male engineers at several of the world’s leading tech companies—but it can be reprogrammed to work better for all humans. Or the policy that gives women 26 times more parental leave than men to take care of a child, reinforcing gender stereotypes of women as carers and men as earners. That policy was created by legislators in the UK—but it can be changed to give every- one an equal opportunity to contribute at work and at home. Or the hiring practices used in nearly 1,500 German companies that resulted in a woman with a Turkish-sounding name and a headscarf needing to send almost five times as many applications as an otherwise identical candidate with a German-sounding name and no headscarf to receive the same number of interview invitations. These hiring practices, too, can be redesigned to eliminate bias.

People’s actions change when the systems surrounding them change—even if their unconscious or conscious biases do not. Not to mention that a single process or policy is much easier to tackle than dozens, hundreds, or thousands of individual minds. By designing smart systems, we can help ourselves live up to our best intentions and perform even better than we could if left to our own devices.

While it would be wonderful to get everyone to change their hearts and minds and shed their unconscious and conscious biases, evidence simply doesn’t suggest that this is a realistic aspiration. Fortunately, we do not need to de-bias ourselves or others to design workplaces that are less biased and more fair.

Sign up for the Big Think Business newsletter
Learn from the world’s biggest business thinkers

Unlock potential in your business

Learn how Big Think+ can empower your people.
Request a Demo

Related