Immersive Learning Reflections from FAT*, Equal Justice Initiative Legacy Museum & The National Memorial for Peace and Justice

The Partnership on AI (PAI) is a non-profit organization dedicated to the study and formulation of best practices on AI, advancing the public’s understanding of AI, and providing a platform for open collaboration between all those involved in, and affected by, the development and deployment of AI technologies. Over the course of this year, we will be piloting a series of “Immersive Learning” experiences with the organizations we work closely with in an attempt to bring technologists and other experts into closer proximity with the contexts and communities that their work ultimately impacts, across a range of domains. We look forward to inviting more guest reflections as a part of this series in 2019.

At the Partnership on AI, we seek to build meaningful engagements between experts in the field – including communities historically marginalized from technology design and deployment decisions. At the recent Association for Computing Machinery Conference on Fairness, Accountability, and Transparency (ACM FAT*) hosted in Atlanta, Georgia, we took advantage of an opportunity to deliver on this objective.

At FAT*, many from across the PAI Partner community presented novel research and were active participants in thoughtful and rigorous discussions around the challenges of incorporating fairness, accountability, and transparency principles into practice. Throughout the conference, conversations about differentiating between bias and fairness in machine learning and of value-sensitive design also related to work by the PAI community to explore these and related issues. The presented research and ensuing conversations raised real-world implications of fairness, accountability, and transparency in daily life, especially for those who are historically marginalized by bias, whether in hiring, education, or encounters with the criminal justice system.

Thematically building off of our participation at FAT*, I was honored to lead an immersive learning and community-building experience with our Partners to the nearby Equal Justice Initiative Legacy Museum, and the National Memorial for Peace and Justice in Montgomery, Alabama. In my role as the Director of Partnerships, I look for opportunities to help bring PAI members into closer proximity with the contexts and consequences of their work. Through this experience, PAI staff and members of the PAI partner community considered the importance of fair, accountable, and transparent machine learning/artificial intelligence technologies while in an environment that draws a clear through-line from the institution of slavery to present-day mass-incarceration in the United States.

Most striking to me were the clear parallels between conditions slaves experienced and those who are currently incarcerated. At the entrance to the museum, visitors are confronted with a video reenactment of family separation, where a young girl is ripped from her father’s neck to be sold away from her family. Three hundred years later along the physical timeline that runs the perimeter of a primary wall of the museum, a photo depicts a fifteen-year-old boy in Florida who has just been sentenced to death: his body pressed up against his mother’s chest, tears of anguish running down both of their faces. Just as these images of injustice echo one another, so too does a video documenting the conditions at Saint Clair Federal Prison depict scenes that recall earlier accounts of the conditions men and women were subjected to during slavery. A clear reminder that the same system of racial inequality that undergirds prisons is still alive and well in many places. This fact that reads along a wall towards the exit of the Legacy Museum: The Alabama state constitution still prohibits racial integration. This has been upheld through two recent attempts at a referendum.

Bryan Stevenson, the founder of the Equal Justice Initiative, has famously said, “First you must tell the truth and then you can get to reconciliation.” The National Memorial for Peace and Justice provides a sobering testament to the 4,400 African American men, women, and children who have been lynched and murdered by mobs between 1877 and 1950. Letters affixed to cast concrete formed to recall weather-beaten wood, read “Thousands of African Americans are unknown victims of racial terror lynchings whose deaths cannot be documented, many of whose names will never be known. They are all honored here.”

We reflected deeply on the dynamic connection between the tragic history of racial injustice in this country with the ways technologies may or may not perpetuate that past. Through these reflections, we resolved that it is in our collective responsibility to see that history doesn’t repeat itself. Partners shared personal motivations ranging from a desire to know more about America’s history to feeling a compulsion to bear witness and internalize the importance of keeping humans centered in our work as technologists.

Seeing the vitality of the modern Montgomery community also bore testament to the resilience of the human spirit, and served as a reminder of the role of community in reckoning with our shared past and building a shared future. The day of learning and reflection closed with lunch at the Kress building, where the local community is leading an urban revitalization effort to preserve the history of the city while celebrating a new generation of artists, business owners, and storytellers that call Montgomery home.I invite our Partners to reflect on the implications revealed through our community-building experiences and to bring those reflections into practice through our work and Working Groups. I look forward to fostering future immersive learning experiences for our partners who were unable to join us in this opportunity; and I ask the extended AI community to consider joining PAI in delivering technologies which promote and reflect fairness, accountability, and transparency.