On October 26, 2020, Walter Wallace, a Black man living with a mental health condition, was murdered by the Philadelphia police. This occurred just one day before Microsoft’s AI for Accessibility Workshop on Societal Bias, Mental Health, and Black Communities. As a gay Black man with a Ph.D. and affiliation with an Ivy League university, I had to muster up enough cognitive resources to have critical conversations about race, technology, and mental health during the workshop as it was all too close to home.

Mental Health America captured my most immediate feelings summarizing that “Processing and dealing with layers of individual trauma on top of new mass traumas from Covid-19, police brutality and its fetishization in the new media adds compounding layers of complexity for individuals to responsibly manage.”

While sitting with these very raw emotions, it is also important to underscore how research and data affect how we understand issues of race and racism. My colleague Dr. Courtney Cogburn suggests that “so much of what we understand about mental health and human behavior is based on White people.”

The workshop on Black people, AI, and mental health sought to center the experiences of Black people while engaging in a hyper inclusive engagement model that brought together individuals from various disciplines to include social work, medicine, AI, human-centered design, and accessibility to discuss the gaps, opportunities, and promise of this work and Microsoft’s role.

How we got here

In May 2020, I connected with Wendy Chisholm, a principal accessibility architect for Microsoft’s AI for Accessibility program, where I learned that Microsoft’s AI for Accessibility team was deeply interested in the need for breakthrough innovation in the mental health space, especially in the midst of Covid-19. The team was eager to incorporate and focus on intersectionality within their mental health projects but was surprised they did not receive proposals from underrepresented groups that addressed mental health and racism. Wendy reconnected with me and noted a particular interest in how to ensure traditionally underrepresented groups show up in mental health data sets that might be analyzed using artificial intelligence. I was then invited to a larger group meeting with Microsoft employees to discuss strategizing how to work with HBCU’s and non-profit organizations to confront the challenge of representation head-on. At the time I was a visiting researcher with the Social Media Collective at Microsoft Research New England and at the very beginning stages of writing a book about Black youth, social media, mental health, and AI, so these topics were top of mind for me. Wendy assembled a diverse and interdisciplinary team to include members from across Microsoft, and then me, a social worker and professor from Columbia University.

What I appreciated the most about that first meeting with the team was the thoughtful and frankly vulnerable understanding of the problem. I took this time to highlight the vast amount of rigorous research happening among Black scholars, in mental health and AI space, that is often overlooked because those researchers were in different networks; social workers talking to social workers, psychiatrist talking to a psychiatrist, and so forth. We also had quite the reflexive conversation about what research was valued and who might be deemed an “expert” on these topics. We all agreed that we needed a convening that disrupted those networks, bringing folks together who may not usually be in the same room, to have critical and robust conversations about Black mental health, AI, racism, and data.

We designed a two-day workshop that would frame the problem as one of structural racism and then amplified relevant anti-racist work happening in mental health to further ground how racism may be suffused in the application of AI for mental health research. Before the conference, I co-designed a pre-survey to identify themes that we might focus on for deeper discussions in breakout rooms. We created three breakout spaces that focused on 1) understanding the landscape of mental health in Black communities, 2) Designing AI for mental health with attention to opportunities and challenges, 3) Designing interventions and interdisciplinary collaborations. The workshop was structured so that there would be a day in between the core content days to quell Zoom fatigue and create a relaxed networking opportunity for invited attendees.

The workshop took place on October 27 and 29 with about 25 attendees. Please visit our website for some of the highlights from our keynotes and breakout discussions.

I was most excited about the conversations that focused on what Microsoft should do next. Where should Microsoft invest its vast resources? A few ideas:

● Develop an AI course that helps researchers and practitioners understand the potential and limitations of AI
● Engage in interdisciplinary praxis- leveraging the voices of social workers, clinicians, researchers, and community members.
● Funding. Invest in supporting people’s time so that they can have time away from their everyday work to more deeply examine issues of AI and mental health in the Black community.
● Co-create. Develop a fellowship that brings diverse voices together to engage in rigorous study, learning, and outputs that lead to innovative research and practice.

The workshop resulted in a report that outlines and summarizes the need to tackle racism in AI and mental health research while uncovering gaps in skill sets, inclusion, and collaboration that can certainly strengthen these critical areas of research.

There is much work to be done and we can only make progress by working together. We must center the voices of marginalized communities, advocate for resources and bring diverse voices to the table. This journey must continue because our lives depend on it. We will be limited in AI innovation and disruption if we do not act now. I am hopeful we can create equitable solutions for everyone. Microsoft is committed to continuing this journey with you. Nothing about us, without us.

Please check out our report and video lectures from the workshop.

Desmond Upton Patton, Ph.D., M.S.W. is an associate professor of social work at Columbia University’s School of Social Work and associate director of diversity, equity and inclusion for the Data Science Institute.