Are you curious about how bias can show up in AI? Do you ever wonder if tools like chatbots and large language models (LLMs) might unintentionally reproduce and amplify biases from the data they were trained on? Join us for an informal workshop where we’ll explore these questions together.

We’ll introduce a simple tool that lets everyone test an LLM under various settings to see how it responds in different scenarios. Then, we’ll brainstorm how we can use this tool to uncover and understand potential biases—even in models that have extensive guardrails. Together, we’ll also discuss ways to design counter-demos that highlight and address risks of racial and gender bias in AI products, which can sometimes be hidden in complex models.

This workshop is a chance to dive into these issues in real time, sharing our ideas and insights as we go. Come ready to experiment, question, and explore with us!

This is a hybrid workshop - whether you will be attending in person or via Zoom, please register at: https://sjsu.zoom.us/meeting/register/tZAucu6urDovHtU18an9BFcTLzRfqGAkCC3p