Stay 9 Seconds:
Go Deeper
Thank you for spending this time with me.
If something in the talk stayed with you, these are some books, films, and studies I'd point you to next.
The through-line of everything below is the idea at the heart of the talk: the avatar is the output. The caricature is the input.
These resources trace that input: where it comes from, how it travels, and why it keeps resurfacing in our newest tools.
The history beneath the avatar
Darkology: Blackface and the American Way of Entertainment — by Rhae Lynn Barnes.
The definitive recent history of how blackface minstrelsy became one of America's most popular forms of entertainment — and how it embedded itself into everyday life for over a century. Essential for understanding why these caricatures were never fringe.
Ethnic Notions — Marlon Riggs (documentary, 1987).
A landmark film tracing the deep-rooted stereotypes of Black Americans — the Mammy, the Coon, the Uncle, the Brute — across two centuries of popular culture. The single best visual companion to the lineage I walk through in the talk.
Toms, Coons, Mulattoes, Mammies, and Bucks — Donald Bogle.
The foundational study of how American film organized Black actors into a small set of recurring types. Bogle gave these figures their names; this is where the "devoted elder" I trace to Grandpa Brian was first catalogued.
Race, data, and AI
Race After Technology: Abolitionist Tools for the New Jim Code — Ruha Benjamin.
A field-defining sociological account of how racism gets embedded in the logic of everyday technology, not as a glitch, but as a default. The framework for understanding why bias survives even when no one intends it.
Artificial Unintelligence: How Computers Misunderstand the World — Meredith Broussard.
A clear, non-technical guide to what AI systems actually do — and why we should never assume the computer gets it right. The best starting point if you want to understand the machinery without a technical background.
“Easily Accessible Text-to-Image Generation Amplifies Demographic Stereotypes at Large Scale” — Bianchi et al. (ACM FAccT, 2023).
The study behind the "attractive," "thug," and "exotic" findings in the talk. It shows generative AI models don't just reflect bias — they amplify it past what exists in the world.Read it →
“Spectacularized and Branded Digital (Re)presentations of Black People and Blackness” — Francesca Sobande.
Scholarship on how digital and AI personas package racial identity as a marketable aesthetic — the academic frame underneath the Shudu and Eva examples.Read it →
On advertising, race, and the "market"
Total Market American: Race, Data, and Advertising — Marcel Rosa-Salas.
My own book. It examines how the advertising and marketing industries don't just segment American consumers by race — they actively produce racial categories in the process. The deeper argument behind much of what I shared today. Read the introduction→
The three questions
If you take one thing from the talk, take these. Three questions to carry into the moment an AI avatar enters your work — or you're asked to build one:
1 — Who is being replaced here, and why? Name the human this synthetic person stands in for.
2 — What's being rehearsed in the story we're telling? Read the profile like the brief it is — and ask where each detail came from.
3 — Why did this character come so easily? That ease is telling. When it arrives without effort, stop — and decide whether or not you will take what culture has offered.