Just weeks after its launch, OpenAI’s new AI video generator Sora is running into some serious trouble.
The online platform, which can create jarringly realistic videos from text prompts, is being swamped with fetish content, and there is growing criticism about its handling of people’s likenesses, particularly when those faces belong to real people.
The core issue? Sora lets users create videos with any face they claim to have permission to use. That sounds fair in theory, but has resulted in a raft of sexually suggestive and fetish-oriented content that’s making many users, especially wome,n think twice about participating.
Why Fears Over Deepfake Fetish Content Are Keeping Women and Hollywood Away from Sora?
Users have found ways to make all sorts of fetish videos using Sora, including foot-fetish clips and more unusual offerings, such as belly inflation videos. While OpenAI officially bans explicit sexual content and nudity, the rules apparently leave plenty of gray area for suggestive material that doesn’t quite cross the line into outright pornography.
The real concern hits when such videos feature the faces of real people. Women who might otherwise be interested in trying Sora have been vocal about their fears.
Many say the likelihood of someone creating creepy or unwanted content with their likeness feels “extremely high,” and that risk alone is enough to keep them away from the platform entirely.
“The trust factor just isn’t there,” one potential user in online discussions explained. “Why would I put my face on a platform where anyone could turn me into the star of their weird fantasy video?”
The backlash isn’t limited to everyday users. Major Hollywood studios and talent agencies are raising alarm bells over Sora’s potential for likeness exploitation, concerned that unauthorized recreations of their clients could appear in videos that might hurt reputations or violate personality rights.
One of the most striking cases involved Fred Rogers Productions, which asked OpenAI to block Mister Rogers’ likeness after videos started circulating that were the complete opposite of the message the popular children’s television host gave to kids.
That even someone as wholesome as Fred Rogers can’t avoid abuse is a sign of how pervasive the problem has become.
Talent agencies are now calling for stricter opt-out controls and stronger legal safeguards. They say the invite-only beta status of Sora simply doesn’t cut it, that protections need to be robust before the platform scales up any further.
Why Sora’s Challenges Exemplify the AI Video Generation Dilemma
To OpenAI’s credit, the company isn’t ignoring the criticism. They’re rolling out new parental controls designed to restrict how different age groups interact within the app, though some details remain murky. For instance, it’s not clear whether adults will still be able to use teen faces for cameo appearances or how crossover interactions will be monitored.
The company is also working on improved moderation tools, more robust content filters, and possible watermarking features to help identify AI-generated content. According to reports, CEO Sam Altman has even suggested that OpenAI might shut Sora down if the harm to society becomes too severe remarkable admission that shows just how seriously they’re taking the concerns.
Sora’s problems really are just one piece of a much larger puzzle. The platform exemplifies the challenges facing all AI video generation tools: the more realistic and accessible the technology, the greater the potential for it to be misused.
Beyond fetish content, experts are concerned about deepfakes being used for impersonation, financial scams, political misinformation, and psychological manipulation. What happens with journalism, legal evidence, or even personal relationships when you can’t be sure a video is real?
The online debate has been heated, with some arguing that the creation of adult content should be legal and taxed as a way to regulate the space, while others demand that platforms take a much harder line in terms of moderation. The right balance between creative freedom and preventing harm is proving incredibly hard to find.
OpenAI’s Sora Challenge, Rethinking Likeness and Trust in the Age of Powerful AI
OpenAI is at a crossroads. The company needs to prove it can manage Sora responsibly before any wider public release. That means not just tweaking content filters, but fundamentally rethinking how people’s likenesses can be used and protected on the platform.
For now, Sora’s “pervert problem,” as some critics have termed it, is a cautionary tale of launching powerful AI tools before the guardrails are set in place. “The technology moved faster than our ability to manage its consequences,” said one industry observer.
Whether OpenAI can rebuild trust and solve these issues remains to be seen. But one thing is clear: the stakes are too high to get this wrong.
