OpenAI’s new text-to-video tool, Sora, has become a viral sensation, reaching over one million downloads in less than five days, according to the company. That pace reportedly outstripped the early performance of ChatGPT when it launched on iOS.
The app, currently invite-only and limited to North America, lets users generate short, realistic video clips by typing simple text prompts. It has soared to the top of Apple’s App Store charts and flooded social media with AI-generated videos ranging from playful to controversial.
When deepfakes cross the line
Among the thousands of videos shared, some depict deceased public figures, including musicians Michael Jackson and Tupac Shakur. Others have recreated scenes involving late comedian Robin Williams, prompting his daughter, Zelda Williams, to publicly plead for users to stop sharing AI-generated clips of her father.
An OpenAI spokesperson told Axios that there were “strong free speech interests” in allowing portrayals of historical figures, but said the company would let authorised representatives of “recently deceased” individuals request removals, though it has not defined how recent that means.
Copyright and corporate tension
Sora has also drawn attention for AI-generated clips of famous fictional characters and brands. Videos circulating online feature characters from Pokémon, SpongeBob SquarePants, The Simpsons and even the McDonald’s mascot Ronald McDonald in surreal or comedic scenarios.
In one popular video, OpenAI CEO Sam Altman appears with Pokémon characters joking, “I hope Nintendo doesn’t sue us.” While Nintendo has not commented, legal experts say such clips could breach copyright law.
“Many of these videos are likely to infringe copyright,” said Professor Mark Lemley of Stanford Law School. “OpenAI is opening itself up to quite a lot of lawsuits by doing this.”
Major studios including Disney and Universal have already filed separate lawsuits against other AI firms for using copyrighted material without permission, and some have warned OpenAI directly against unauthorised use of their intellectual property.
OpenAI’s response
OpenAI says it is working with rights holders to manage these concerns. In a recent blog post, CEO Sam Altman said the company plans to give creators and studios “more granular control” over how their characters appear in user videos. He also suggested a potential revenue-sharing model for approved content in the future.
“We’re learning quickly from how people are using Sora,” Altman wrote. “We’ll give rights holders more control and share revenue where appropriate.”
The company has also built tools for users to flag copyright or likeness violations and allows individuals to manage how their own image, or “cameo”, is used within the app.
A powerful but controversial tool
Sora’s technology represents a huge leap forward for AI-generated media. In seconds, anyone can create lifelike video scenes from text, a process that once required teams of artists and animators.
But the same technology has also fuelled fears about deepfakes, misinformation, and digital impersonation. As AI video becomes indistinguishable from reality, questions of authorship, consent, and truth are increasingly pressing.
For now, Sora’s success captures both the promise and peril of generative AI: a new form of creativity that could redefine entertainment or force the tech industry into yet another round of courtroom battles.
