OpenAI’s New App Sora Sparks Debate Over AI-Generated Fake Videos 

OpenAI criticized after Sora deepfakes show MLK Jr. in disrespectful AI-generated videos.

SAN FRANCISCO — Security footage of a tech CEO shoplifting, Ronald McDonald in a police chase, and even a selfie-style clip of Jesus joking about “last supper vibes.” None of it was real. Instead, these AI-generated clips were among the most popular posts Wednesday on Sora, a new TikTok-style app from OpenAI that allows anyone to create hyper-realistic fake videos.

Launched this week in the U.S. and Canada, Sora uses OpenAI’s advanced video-generation technology to let users build short clips of themselves, friends, or entirely fictional scenes with striking realism and matching audio. Within 24 hours of release, users flooded the platform with everything from comedy skits to manipulated videos of real people, raising new concerns over deepfakes and misinformation.

Early tests showed the tool could recreate police body-cam footage, TV shows, and even convincing clips of historic figures like John F. Kennedy. Experts warn that the technology could make it increasingly difficult to distinguish authentic footage from fabricated content.

“The challenge with tools like Sora is that it makes the problem exponentially bigger — it’s so available and so good,” said Ben Colman, CEO of Reality Defender, a firm that develops software to detect deepfakes.

OpenAI emphasized that Sora includes safeguards, such as banning impersonation, scams, nudity, and violence. The company says users control how their likeness is shared through a “cameo” feature, which lets them appear in AI-generated videos made by friends or other users. They can also delete videos that misuse their image.

Still, users have already found ways to exploit the system. YouTuber Justine Ezarik, known as iJustine, discovered her likeness had been used without her consent in inappropriate videos. Although the account responsible was quickly removed, Ezarik said she closely monitors her presence on the app to delete anything that “crosses a line.”

Copyright concerns are another looming issue. OpenAI said it would remove copyrighted characters if rights holders request takedowns. Yet, the company is already facing multiple lawsuits over how it trained its AI models, including allegations of using copyrighted material without permission.

Despite these challenges, Sora surged to the third-most-downloaded app on Apple’s App Store just one day after its debut, even though access currently requires an invite. Competitors like Google and Meta have also launched AI video tools, but OpenAI is the first major tech firm to build a full social platform around them.

Industry watchers predict Sora will force regulators, platforms, and creators to confront the growing influence of AI video. “Even if viewers prefer authenticity, if it becomes hard to discern, then it won’t matter much,” said Mathieu Samson, founder of AI filmmaking agency Kickflix.

For now, Sora is reshaping online video — raising questions about creativity, consent, and the blurry line between imagination and reality.