Scrolling through the Sora app feels like entering a surreal multiverse where absurd and hyperreal short videos abound: Michael Jackson performing standup, the Predator alien flipping burgers, a moose crashing through a glass door, and Queen Elizabeth diving off a pub table. Released by OpenAI, the creator of ChatGPT, Sora generates entirely AI-crafted clips that mesmerize but also force users to constantly question the authenticity of what they see. Sam Gregory, a deepfake expert and human rights advocate, warns that Sora makes it nearly impossible to distinguish real from synthetic content, generating a pervasive “fog of doubt” over all media. Despite featuring only AI-generated videos, the app swiftly went viral, surpassing a million downloads in under a week after its September 30 launch, topping the U. S. App Store. Currently, Sora is available exclusively to U. S. iOS users with invite codes. To create videos, users scan their face and record a voice sample, then enter custom text prompts to generate 10-second hyperreal clips complete with sound and dialogue. The “Cameos” feature allows superimposing a user’s or friend’s face onto existing videos. Although Sora watermarks videos, many websites now remove these marks. Initially, OpenAI allowed default recreation of copyrighted material unless owners opted out, leading users to generate AI content featuring characters from “SpongeBob SquarePants, ” “South Park, ” “Breaking Bad, ” “The Price Is Right, ” and “Friends. ” This escalated to re-creating deceased celebrities such as Tupac Shakur wandering Cuba’s streets, Hitler confronting Michael Jackson, and Martin Luther King Jr. delivering altered speeches demanding the release of rapper Diddy. Zelda Williams, daughter of Robin Williams, condemned these videos as disrespectful, expressing anger at the exploitation of human lives and artistic legacies for online likes. Other AI-generated depictions of Kobe Bryant, Stephen Hawking, and President Kennedy have gone viral across social platforms. Fred Rogers’ family expressed frustration over misleading Sora videos portraying Mister Rogers holding a gun or interacting with Tupac, contradicting his genuine child-centered values.

They have formally requested OpenAI to block his voice and likeness on the platform, expecting respect for personal identities moving forward. Similarly, Hollywood talent agencies and unions like SAG-AFTRA criticize OpenAI for unauthorized use of actor likenesses and licensed characters without fair compensation. Responding to copyright concerns, OpenAI CEO Sam Altman announced plans to give rights holders greater control over how their characters appear in AI videos and explore revenue sharing. Studios can now “opt-in” to allow character use—a reversal of the initial opt-out policy. Altman envisions a future where personalized content thrives, likened to an explosion of creativity akin to “interactive fan fiction. ” Meanwhile, estates of deceased actors race to protect their likenesses amid AI advances. CMG Worldwide partnered with deepfake detector Loti AI to monitor unauthorized digital recreations of 20 celebrities, including Burt Reynolds and Rosa Parks. Loti AI reports a 30-fold increase in signups since Sora 2’s launch as users seek to regain control over digital images, having removed thousands of unauthorized deepfake instances since January. Following disrespectful depictions of Martin Luther King Jr. , OpenAI paused his image generation per the estate’s request, acknowledging that while free speech permits portrayals of public figures, their representatives should control usage of their likenesses. Now, estates can request blocking of their likeness from Sora’s Cameo feature. Legal pressures have prompted Sora to tighten rules on copyrighted content, issuing violation warnings for recreations like Disney characters. Some users mock the restrictions by creating memes about these content policy notices. Meanwhile, “AI slop” — bizarre, viral content such as grandmothers chasing crocodiles or comical “fat olympics” — dominates engagement, with dedicated creators turning it into a lucrative stream dubbed “Cocomelon for adults. ” Critics warn that casual appropriation of anyone’s likeness, not just celebrities, risks public confusion, misinformation, and erosion of trust. Furthermore, the technology’s abuse by bad actors and governments for propaganda heightens concerns. Gregory fears fabricated protest footage, staged atrocities, or manipulated videos placing false words in real people’s mouths could severely damage truth and accountability amid the flood of AI-generated content.