- Katica Roy
How to Catch Unconscious Bias: a 4-Step Framework
Updated: Dec 14, 2020
#algorithms #artificialintelligence #implicitbias #unconsciousbias
Welcome to my weekly Q&A roundup. (Scroll down to find the Q&A.)
If this is your first time here, welcome. I spend a fair amount of time speaking at events and conferences. At the end of my presentations, I leave space for audience members to ask questions—tough questions, brave questions, you name it.
The level of candor and curiosity always inspires me, and I want to share that sentiment with you. So each week, I pick one question that I believe others would find most instructive and publish my response to it here.
The purpose of this weekly tradition is transparency and inclusivity.
• Transparency: a behind-the-scenes look at my day-to-day.
• Inclusivity: bringing others along in the journey.
I Catch My Biases with This 4-Step Framework
The unconscious brain processes 200,000 times more information per second than the conscious mind. Our brains rely on patterns (or shortcuts) to process all this information efficiently, and that leaves room for unconscious bias to sneak in.
How can we catch and overcome our unconscious biases?
Mainstream “solutions” for catching individual biases range from online testing to in-person training and nearly everything in between.
I like to use the following 4-step framework to catch my biases. I prefer this framework because it’s a data-driven approach to address unfiltered thoughts, doubts, and curiosities.
A quick note before we get started: my framework is more involved than most implicit bias training/activities—and, as a result, it has opened huge doors and remarkable opportunities for growth. It’s worth the investment.
1. Start with a hypothesis.
We live with an unlimited stream of stories, news, and information. As we take in these messages, it’s tempting to want to make sense of them by jumping to a conclusion about what they might mean. Instead, I allow myself to be uncomfortable with not having a clear conclusion. I choose to leave an open-loop and be ok with uncertainty. Researchers call this a hypothesis:
Here’s what I believe is true based on what I know now, and I accept that my theory is the starting point for further investigation.
2. Dive into the data.
As a result of living with nearly unlimited access to information, we must exercise rigor and diligence in step two. As I dive into the data to test my hypothesis, I carefully vet all sources and approach my hypothesis from as many angles as possible to avoid confirmation bias. This means seeking out data that may disconfirm the hypothesis.
3. Follow the data.
I’ve said that most people use data like an inebriated person uses a lamppost. That is, they use the lamppost for support instead of illumination. To ensure that I use data correctly—for illumination and not for support—I go to where the data leads. And where the data leads is not necessarily where I want the data to go. Again, I allow myself to be uncomfortable here.
4. Share what the data says.
Publicly sharing my findings necessitates transparency (#bulletproof). Others, especially those whose viewpoints may conflict with my findings, are welcome to follow the breadcrumbs that led to my conclusions.
Using this framework has not only made me a stronger intellectual, it’s also given me a leg up in exploring uncharted territory—discovering fresh insights and answers to some of the world’s toughest problems.
I believe technology can be used for good. As such, I believe unbiased algorithms can scale this framework across industries to create a more equitable world.
These Q&A roundups can be delivered directly to you—a week before I publish them here. Interested?
Join the Brave Souls® community
(All you need is an email address.)