09/28/2026
Ask an AI chatbot for help writing a work email. Use the word "maybe." Say "I think" or "we." Sign it with your name.
Now ask the same thing without the hedges.
Johns Hopkins researchers found that four major AI models gave back less formal, less sophisticated responses when the prompt used language patterns more commonly associated with women. Not because of the name attached... because of how it was asked.
The bias is baked into how these tools read us. And most of us have no idea it's happening.
This is the kind of thing we're digging into at WF Presents 2026. Dr. Meredith Broussard, data journalist, NYU researcher, and author of "More Than a Glitch," is joining us on November 18 to talk about the systems we use every day, and what's really going on inside them.
Reservations are open. Early bird pricing ends October 1 >> https://womensfundmke.org/wf-presents-ai