OpenAI Halts Martin Luther King Jr. Sora Videos After ‘Disrespectful’ Deepfakes Spark Outrage

OpenAI has officially "paused" the ability for its text-to-video model, Sora, to generate videos depicting Dr. Martin Luther King Jr. This decision comes after a wave of "disrespectful" and racist AI-generated videos of the late civil rights icon flooded the internet, sparking outrage from the public and a direct request from the King family.

The controversy erupted shortly after the launch of Sora's newest version, which allows users to create hyper-realistic "cameos" of historical figures. Users began creating deeply offensive content, including videos reportedly showing Dr. King making monkey noises during his "I Have a Dream" speech, wrestling Malcolm X, stealing from a grocery store, and perpetuating racist stereotypes.

A "Gross" Misuse of Technology
The backlash was swift and severe. Dr. Bernice King, the daughter of Dr. King, took to social media to demand the videos stop, concurring with a similar plea from Zelda Williams, daughter of the late actor Robin Williams, who had condemned AI-generated videos of her father as "gross."

In a joint statement released on Thursday, OpenAI and the Estate of Martin Luther King, Jr., Inc. (King, Inc.) addressed the issue directly.

"Some users generated disrespectful depictions of Dr. King's image," the statement read. "So at King, Inc.'s request, OpenAI has paused generations depicting Dr. King as it strengthens guardrails for historical figures."

A New Precedent for AI Guardrails
The incident has forced a major policy shift at OpenAI. The company, which had previously allowed the generation of historical figures, is now grappling with the ethical and legal nightmare of its powerful technology being used to create and spread hateful content.

While acknowledging the "strong free speech interests in depicting historical figures," OpenAI's statement conceded that "public figures and their families should ultimately have control over how their likeness is used."

To that end, OpenAI has now created a new policy allowing authorized representatives or estate owners of a public figure to request that their likeness not be used in Sora. This is a significant move as the company scrambles to build safety guardrails for a technology that many fear is evolving faster than society's ability to control it.