AI-generated images often reflect bias and exaggerate stereotypes related to race, gender, age, and disability. I presented at the Accessible and Inclusive Design Conference earlier this month about minimizing bias in AI-generated images. This conference was a free virtual event focused on creating and building accessible and inclusive learning environments. I presented techniques for writing more effective image prompts to minimize bias and reflect more inclusive representation in your learning experiences. I also discussed some of the ongoing challenges with inclusive representation in AI image generation, such as representing visible disabilities. AI image generation gives us the opportunity to create more inclusive images than would be possible with stock and character image libraries—but only if we prompt carefully to counter AI’s inherent biases.
Watch the recording
You can watch the recording of my presentation now. If the video isn’t embedded below, try watching directly on YouTube.
Download the slides
You can also download the slides to get all of the example prompts and the suggested prompt structure I provided in my session.
More on accessible and inclusive design
Check out the recordings of all of the sessions from AIDC to catch up on any sessions you missed. Recordings from the past three years are also available.
In my session, I cited Tristin Sylvester’s Training Mag Network webinar: Ethical AI In Design: Craft Inclusive Learning Beyond Biases. She did such a good job explaining the problem of bias in AI that I actually just shared a few quick examples in my session and then referred people to her.
