I was intrigued and impressed when OpenAI first demonstrated Sora earlier this year. There seemed to be no limit to the movies the AI video model could produce from a text prompt. Sora effortlessly turned descriptions into immersive, realistic videos, and OpenAI coyly hinted at a general rollout in the near future. Months later, only professional filmmakers working with OpenAI have any real access (a recent, short leak doesn’t count). The same goes for other much-hyped AI video generators, including Meta’s Movie Gen and Google’s Veo.
Lots of great AI video generators have emerged since Sora blew people away, but it’s hard not to feel like a kid sitting with their nose against the glass of the toy store wondering why we can’t have a little fun with it can play toys. This is why I think OpenAI and the rest of the cagey AI video creation models are still locked down.
Movie trailers always lead to disappointment
Maybe I’m just a skeptic, but I find it strange that OpenAI, Meta, and Google all seemingly couldn’t wait to show off demos of their respective AI video generators without even a vague idea of a rollout date. It reminds me of movie teaser trailers that come out a year before a movie and promise far more than the final version can deliver. I wonder if Sora, Movie Gen, and Veo might have some cooking to do before we get our hands on them.
The carefully curated demos may not only be the best examples of the AI models, but also the only ones worth showing to the public. Sora’s default output could be more of a fever dream than a heavenly vision. Maybe asking for a “serene sunset over a lake” only occasionally yields a peaceful evening on the water. If nine out of ten Sora clips depict a lake melting into a neon green abyss under a sun that flickers like a ghostly flashlight, I wouldn’t do that. I can’t blame OpenAI for holding Sora back for now.
Ethics (or legal exposure)
The companies behind AI image and video creation tools usually make a point of highlighting their ethical training and output controls wherever possible. Sora is no exception, but the lines of ethical model boundaries become a lot blurrier with videos than with images, especially since the video is essentially a large number of images strung together.
Unapproved data creating deepfakes of real people without their knowledge and producing films with trademarked characters and logos without permission opens the gates of a vast legal and ethical minefield. Working with professional filmmakers and commercial video directors eliminates these issues because the tech company can keep a close eye on the AI’s output and prevent occasional breaches.
Where’s the profit?
As much as OpenAI, Adobe, Google and Meta love to show off their technology, the people who control the wallet want to know where the return on this investment will come from and when. The goal is a polished, marketable AI video generator, not a cool toy. A free AI video playground for experimenting and making mistakes is a step on the road, not the destination.
While we don’t know the exact cost, it’s possible that high-end AI creators are just as expensive to run compared to Runway or Dream Machine. The processing power required is certainly staggering compared to AI text composition. If you scale it up without restrictions, it could cause a server meltdown. Allowing bored students to make short clips of a dog playing the violin in a submarine might not seem like it’s worth keeping Sora running around the clock for millions of users. Restricting access to approved professionals gives companies more control.
OpenAI is almost certainly working on strategies to monetize hobbyists, smaller marketing companies, and movie studios that are willing to pay for continued access to advanced AI video generators like Sora. But until they’re as accessible as the subscriptions for premium versions of ChatGPT, Gemini, and other AI chatbots, only the deepest filmmakers will likely get access to Sora and its sister models. Until then, we are just spectators.