Who we are

We are an interdisciplinary team on a mission to teach the world how to manage algorithmic bias, fairness, transparency, and accountability.

Our story

2019: We have a problem

Bias in machine learning (ML) algorithms---that is, adverse outcomes that affect particular people more than others, particularly historically marginalized people---is widespread and harmful. In 2019, Dr. Nick Merrill identified a problem: students at Berkeley were graduating and getting jobs developing widely-used machine learning algorithms at some of the biggest companies on earth, but never received a rigorous education in algorithmic fairness. If these graduates, among the best-trained engineers in the world, were not equipped to identify bias, we're in trouble.

2020: The MLFailures bootcamp

From this observation, the MLFailures bootcamp was born. Dr. Merrill, along with Dr. Nitin Kohli, Samuel Greenberg, Inderpal Kaur, and Jasmine Zhang, developed a one-week bootcamp for UC Berkeley graduate students. By 2020, these two lectures and labs were educating hundreds of students yearly. These labs slowly spread throughout the country and the world.

2022: Shaping policy & practice

As our higher-education coursework spread, we realized that the desire for high-quality resources for learning about machine learning bias and fairness spread far beyond higher education.

Today, we provide training, education, and to clients including for-profit, non-profit, academic, and government organizations. The Broad Daylight team has educated and advised U.S. government clients including Congressional aides and chiefs of staff, mid- and senior-level Executive Branch officials, the U.S. Postal Service, and corporate clients including Ad Hoc LLC and Unify ID.

Experiential learning

We are an interdisciplinary team on a mission to teach the world how to manage algorithmic bias, fairness, transparency, and accountability. We focus on building internal capacity, teaching tools, and strategies to identify, interrogate and ameliorate algorithmic bias. We deliver training and advising to clients including for-profit, non-profit, academic, and government organizations.

The core of our methodology is experiential learning. People learn by doing. We deliver hands-on learning experiences at all levels of technical detail, from "no-code" interactive labs to full-on coding challenges. All of our work is based on real-world incidents.

Get smart about algorithmic bias

Build your organization's capacity to identify and ameliorate algorithmic bias. Get in touch and let us know how we can help.