Session Spotlight

Rhea Eckman

Camp Counselor

Machine Learning and Bias: Drawing Insights from History to Build a Better Future

Event Logo

Wednesday, July 31, 2024 - 9:00 PM UTC, for 1 hour.

Regular, 60 minute presentation

Room: African 30

Machine Learning
Bias
Ethics
Community

This talk is an invitation to learn about the history of discrimination, activist movements, and legal protections within three fields in which machine learning algorithms have recently been shown to demonstrate similar bias: Hiring Practices, Predictive Policing, and Healthcare Diagnostics. I will also be discussing Google Gemini’s “overcorrection” of racial bias, using this as a case study to dig into some of the more nuanced issues we encounter when tackling problems this large and pervasive. In addition to activists and other people who made their mark on history, I believe that we have much to learn from people in our communities who experience discrimination in their day-to-day lives. So, throughout the presentation I will also be sharing stories from individuals living in the St. Louis region who have experienced the effects of bias, along with their fears and hopes for a future that is increasingly becoming dependent on machine learning. With our technological systems rapidly evolving, it is natural to be excited about the potential to address problems in society that have previously felt unsolvable. It is important, however, that we critically reflect on what kind of future we are working to build, as well as the scope and nature of the problems we are trying to solve. Let’s work to build a better future without repeating mistakes of the past!

Prerequisites

Things to know: No experience required! Things to have: If you have any stories you are willing to share, or news stories you think are worth mentioning around these topics, those would be great to bring. Also, bring an open mind, and a notepaper and pen/pencil if that's your style.

Take Aways

  • Increased appreciation for the many ways in which various individuals and organizations have worked to address the sources and ramifications of bias over time
  • Potential strategies for addressing bias, sourced from recent and old history
  • Increased knowledge about the potential benefits and risks of machine learning
  • Ideas for creative solutions moving forward, leveraging the insights provided by parallel professional sectors as well as that of those operating outside of those realms
favorited by:
Ross Larson Josh Gretz Micha Rodriguez