Filmmaker Valerie Veatch: AI Enthusiasts Overlook the Racism and Eugenics in Their Tools

Documentary filmmaker Valerie Veatch came to generative AI the way most people do: out of curiosity. When OpenAI opened public access to Sora, its text-to-video model, in 2024, she did not fully understand how the technology worked, but she wanted to see what it could do. She was especially drawn to the community of artists who shared their AI work online. It looked like a new space for creative dialogue.

What Veatch found inside shook her. Generative models kept producing images steeped in racism and sexism: stereotyped depictions that reflect deeply rooted biases in the training data. Bias in AI is not news in itself; researchers have been talking about it since the first language models appeared.

But Veatch ran into more than the problem itself. She ran into the reaction to it, or rather the lack of one. The enthusiasts in her community mostly did not notice the problem or chose to ignore it.

The technology was interesting, the results were striking, and nobody especially wanted to deal with the fact that the machine reproduces a historically formed hierarchy. For Veatch, this indifference was a more alarming signal than the artifacts in the AI output.

Her experience ultimately became the basis for the documentary Ghost in the Machine, which explores not so much the technical workings of generative AI as the ideology behind it. The title of her interview with The Verge, "Gen AI Kool-Aid tastes like eugenics," sounds like a provocation, but it carries a specific thesis: the narrative of progress that the industry promotes grows from the same soil as eugenics. The idea that technology will make the world better by automatically weeding out the "inefficient" and the "imperfect" repeats a logic that humanity has already condemned in another context.

Veatch is not the first to draw this parallel. Critics of tech optimism have long pointed out that the Silicon Valley narrative of "improving" people and systems through data and algorithms carries a hierarchical, and at times openly discriminatory, subtext. But the documentary format lets her show this through personal experience rather than academic argument, and that is what makes her position hard to wave away.

Her path into the AI space is typical. Most users first admire the capabilities, then notice the problems, and this is where the road forks. Some start demanding change. Others rationalize: "the technology is neutral," "these are data artifacts," "it will get better over time." Veatch sees this silent acquiescence as the main symptom.

What she describes is not a technical bug that the next update can fix. It is a question of whose values get encoded into the tools that shape visual culture. And while the generative AI industry grows faster than it can come to terms with its own consequences, voices like Veatch's remain an important counterweight, all the more so because they ask their uncomfortable questions from the inside, not the outside.