Stanford’s AI posters replaced a real student with a fictional one

What the poster changed

Billy Ramirez, the student removed from the original photograph, initially called the incident “funny.” He later told Stanford Review, the student newspaper that first reported the mistake, that he felt silenced and excluded from an image intended to reflect the community.

The edits went beyond replacing Ramirez with a fictional Black student. A young Asian woman and a white man in the photograph were made slimmer and closer to conventional ideas of attractiveness.

According to the San Francisco Chronicle, the edited image appeared on “Welcome Home” posters around the university. Ramirez said they had previously been “everywhere,” but staff are now taking them down.

Ramirez said he loves Stanford but dislikes the way the university handled the campaign. The fictional student, he argued, does not reflect reality: the campus is not as diverse as the poster tries to suggest.

This was not Stanford’s first AI poster

The university’s Residential & Dining Enterprises department had already used AI to create dining posters advertising different offers, the San Francisco Chronicle found.

Those images showed food with excessive detail and distorted forms, familiar signs of poor image generation. The earlier posters matter because they suggest the problem was not a single accidental edit, but a broader willingness to use generated imagery in university communications without enough scrutiny.

Stanford says its rules were clear

A Stanford representative confirmed that AI was used to edit the images and that the work violated university policy.

Stanford’s rules strictly prohibit using AI to create or alter images of:

The university said both the photo manipulation and the lack of disclosure violated those rules. It plans additional staff training and says all materials will receive proper review.

I think the central failure here is not that the generated image looked artificial. It is that the image was designed to look authoritative while quietly changing the people and reality it claimed to represent. The result was not just an inaccurate picture, but a polished institutional statement built from invented identity.

What I’d want to know is who approved the posters, what review they received, and why the university’s existing rule did not stop them. Training may reduce repeat mistakes, but it does not explain why a basic check was missing when the image was made.

Ramirez said he was tired and somewhat bewildered by the episode. He was not especially surprised, he added, because this is Silicon Valley and Stanford has an unusually active AI culture. Some of his professors regularly use Anthropic’s Claude chatbot.

That culture may make AI adoption feel ordinary inside the institution. It does not make an invented student ordinary outside it: a university can experiment with tools, but it cannot outsource responsibility for who gets shown as real.