On Tuesday, OpenAI revealed it is discontinuing Sora, a social application resembling TikTok that debuted just six months ago. The company provided no specific reason for the shutdown or a precise date for when the service will officially end. Initially launched as an invite-only network, Sora generated significant buzz, with many eager to secure access. However, similar to Meta’s Horizon Worlds—a virtual reality social platform once central to the company's metaverse ambitions but now facing its own challenges—Sora failed to maintain lasting engagement. Despite the remarkably advanced Sora 2 model for generating video and audio, consistent interest in a social feed powered solely by AI never materialized.

Designed as an AI-centric version of TikTok, Sora replicated the familiar vertical video feed interface. Its standout feature, originally called “cameos,” enabled users to scan their faces to create realistic deepfakes of themselves. These digital avatars could be set as public, allowing others to produce videos featuring them. (The company Cameo successfully sued OpenAI over the feature's name, leading to its rebranding as “characters.”)

Unsurprisingly to many, this advanced deepfake application proved to be deeply unsettling. At launch, Sora often felt like a poorly moderated landscape filled with disturbing videos of Sam Altman. One particularly jarring example featured a hyper-realistic clone of the OpenAI CEO wandering through a facility with overfed pigs, inquiring, “Are my piggies enjoying their slop”—an experience that left a lasting impression.

Although Sora’s policies aimed to prevent the generation of videos featuring public figures without explicit consent, users frequently bypassed OpenAI’s safeguards. This led to the appearance of deepfakes depicting individuals like civil rights icon Martin Luther King, Jr. and actor Robin Williams, prompting their daughters to publicly appeal on Instagram for users to stop creating videos of their late fathers. After producing numerous clips showing Sam Altman humorously stealing Nvidia chips from a Target store, the user base pivoted. They began intentionally generating content with copyrighted characters, seemingly to provoke legal action against the figure they enjoyed deepfaking—resulting in scenes like Mario using marijuana, Naruto ordering Krabby Patties, and Pikachu performing ASMR.

This strategy did not unfold as anticipated. Instead of pursuing litigation, Disney—a company known for vigorously protecting its intellectual property—entered into a $1 billion investment and licensing agreement with OpenAI. The deal would have permitted Sora to generate videos featuring characters from Disney, Marvel, Pixar, and Star Wars, marking what appeared to be a pivotal moment for the AI sector. However, with Sora’s discontinuation, the agreement has also dissolved; notably, it seems no funds were exchanged prior to its collapse. (Disney commented on the situation, telling the Hollywood Reporter it plans to “continue to engage with AI platforms” moving forward.)

The initial excitement surrounding Sora was undeniable. According to data from mobile intelligence firm Appfigures, the app reached its peak in November with approximately 3,332,200 downloads across the iOS App Store and Google Play. Had growth continued, OpenAI might have sustained the service, but that trajectory stalled. By February, downloads fell to about 1,128,700. While this figure may seem substantial, it pales in comparison to ChatGPT’s 900 million weekly active users. Over its lifespan, Appfigures estimates Sora generated roughly $2.1 million from in-app purchases, which allowed users to buy additional video generation credits. Given that OpenAI already operates at a significant loss, it’s unlikely the app’s computational costs were decisive. However, if the platform wasn’t expanding, its potential liabilities may have outweighed its value.

When OpenAI first introduced Sora, it signaled a future where deepfake technology could become easily accessible to everyone. Although I rarely post on TikTok, I felt compelled to share a public service announcement about the rapid arrival of this concerning technology. That video garnered over 300,000 views—far exceeding the usual engagement on my mostly inactive account—highlighting the genuine public concern this news sparked. I never anticipated the app would last only half a year.

Yet, Sora’s disappearance does not mean the underlying risks have vanished. The Sora 2 model remains available, though now hidden behind ChatGPT’s paywall. Moreover, OpenAI is far from the only company making such technology widely accessible. It is only a matter of time before the next social AI video application emerges, potentially flooding the digital landscape with another wave of clips—perhaps even featuring Snow White storming the Capitol.