Problem

For the final deliverable of this series, I wanted to see how AI would visualize race when given a broad prompt about the Asian American experience. One of the visible ways that AI contributes to racism is by using algorithms to construct racialized media. For the purpose of this assignment, I chose to use two generative AI platforms, Storybird.ai and Leonardo.ai, to create a flip through book and a video about Asian American affinity groups.

For instance, AI often perceives professionalism to be associated with white identities.

Process

The tool I used allowed me to input a prompt from my own systems and see how race influenced visual outcomes. The following prompts were put into each platform:

Storybird.ai & Leonardo.ai
"Act like an educator who is dedicated to social justice and uplifting the Asian American experience by facilitating an affinity group. As a facilitator for an Asian American Affinity group, your role is to guide students in reflecting on their race and identity. Introduce yourself and emphasize your coaching role in this reflective process. Follow these steps below to design a storybook that is culturally inclusive and meaningful and discusses the model minority myth. Make sure the book is at least 10 pages long and no longer than 15 pages.

The learning outcomes are to:
1. Make connections across grade levels/school for students
2. Engage in conversations about race and experience
3. Help them understand and foster a sense of community and support"

Results

To view the Storybird.ai book, click here.
To view the Leonardo.ai video, click here.

Implications

One notable aspect of this project I chose to do was to leave Asian American up to AI to interpret. In the storybook, it was clear that AI understood the Asian American experience as East Asian and not encompassing all Asian American identities. It is interesting to see how AI creates culturally inclusive materials by drawing and forming assumptions of race. It was able to understand the Model Minority Myth without having any context of it in the prompt, which suggests that algorithms do provide context, but perhaps in a more narrow scope than we hope.

Questions to consider:

Can AI be disentangled from race if our algorithms are media driven?
What does the future of AI hold if it can only categorically understand Asian American as East Asian?

AI, Race, and Visualizations