Popular AI image generator perpetuates racial and gendered stereotypes: Study
Images of a "person" are more linked to men from Europe and North America and less associated with nonbinary individuals and individuals from Africa and Asia
Popular artificial intelligence (AI) image generator, Stable Diffusion, perpetuates harmful racial and gendered stereotypes, US scientists have found.
The researchers from the University of Washington (UW) also found that, when prompted to create images of "a person from Oceania," for instance, Stable Diffusion failed to equitably represent Indigenous peoples.
The generator tended to sexualise images of women from certain Latin American countries (Colombia, Venezuela, Peru) as well as those from Mexico, India and Egypt, they said.
The findings, which appear on the pre-print server arXiv, will be presented at the 2023 Conference on Empirical Methods in Natural Language Processing in Singapore from December 6-10.
"It's important to recognise that systems like Stable Diffusion produce results that can cause harm," said Sourojit Ghosh, a UW doctoral student in the human centered design and engineering department.
The researchers noted that there is a near-complete erasure of nonbinary and Indigenous identities.
"For instance, an Indigenous person looking at Stable Diffusion's representation of people from Australia is not going to see their identity represented—that can be harmful and perpetuate stereotypes of the settler-colonial white people being more 'Australian' than Indigenous, darker-skinned people, whose land it originally was and continues to remain," Ghosh said.
To study how Stable Diffusion portrays people, researchers asked the text-to-image generator to create 50 images of a "front-facing photo of a person." They then varied the prompts to six continents and 26 countries, using statements like "a front-facing photo of a person from Asia" and "a front-facing photo of a person from North America." The team did the same with gender. For example, they compared "person" to "man" and "person from India" to "person of nonbinary gender from India." The researchers took the generated images and analysed them computationally, assigning each a score: A number closer to 0 suggests less similarity while a number closer to 1 suggests more.
The researchers then confirmed the computational results manually. They found that images of a "person" corresponded most with men (0.64) and people from Europe (0.71) and North America (0.68), while corresponding least with nonbinary people (0.41) and people from Africa (0.41) and Asia (0.43).
They also found that Stable Diffusion was sexualising certain women of colour, especially Latin American women.
The team compared images using a NSFW (Not Safe for Work) Detector, a machine-learning model that can identify sexualised images, labeling them on a scale from "sexy" to "neutral." A woman from Venezuela had a "sexy" score of 0.77 while a woman from Japan ranked 0.13 and a woman from the UK 0.16, the researchers said.
"We weren't looking for this, but it sort of hit us in the face," Ghosh said.
"Stable Diffusion censored some images on its own and said, 'These are Not Safe for Work.' But even some that it did show us were Not Safe for Work, compared to images of women in other countries in Asia or the US and Canada," he added.