A recent Asian American Massachusetts Institute of Technology (MIT) graduate made headlines this month after sharing evidence of racial bias in artificial intelligence (AI).
What she asked for: Rona Wang, 24, has been experimenting with AI portrait generators. In a recent attempt, she purportedly used a program called Playground AI to turn an image of her in an MIT sweatshirt into “a professional LinkedIn profile photo.”
What she got: The program, to Wang’s surprise, returned an image that gave her a fairer complexion, dark blonde hair and blue eyes.
“I was like, ‘Wow, does this thing think I should become white to become more professional?’” she told Boston.com.
Reactions: On July 14, Wang shared the result in an X post, which has now received over 5.9 million views.
Social media users replied to the post with their own takes and theories. Druv Bhagavan, a M.D./Ph.D. student at Washington University School of Medicine in St. Louis, warned about delegating important tasks to AI tools as “unbiased training data is rarely ever used.” Meanwhile, user @CityBureaucrat wrote, “AI stands for Aryan Intelligence.”
Playground AI’s response: Responding to Wang’s post, Playground AI founder Suhail Doshi said that “models aren’t instructable like that” and will pick “any generic thing based on the prompt.” However, he said in another tweet that Playground AI is “quite displeased with this and hope to solve it.”
The big picture: Racial bias is an ongoing concern in AI-generated images. While some programs have been accused of turning subjects white, others have been slammed for turning subjects Asian. Outside image generation, racial discrimination in AI facial recognition is also a concern, particularly among African Americans. How and whether artificial intelligence manages to solve these issues are yet to be seen.