An AI Movie Made in One Text Prompt
Are you ready for a world where viewers can instantly become creators?
“[You will] be able to render a whole movie. You’ll just come up with an idea for a movie, then [an AI chatbot] will write the script, then you’ll input the script into the video generator and it will generate the movie. Instead of watching some movie that somebody else came up with, I can just generate and then watch my own movie.”
Ashton Kutcher said this last month at a Berggruen Salon after the actor and venture capitalist played around with the beta of Sora, OpenAI’s new text-to-video model. (This is a sentence I never thought I’d type after watching Dude, Where’s My Car?)
Kutcher’s vision is a common one among both VCs and members of the AI Twitter community, especially since OpenAI announced Sora in February. Although Sora is not yet available to the public, it does appear to be super impressive. It’s able to generate realistic-looking videos up to 60 seconds long from just a text prompt, and filmmakers who have gained access to Sora have had their AI films showcased at the Tribeca Festival, among other places. It’s also already being used in commercial productions such as this recent Toys “R” Us ad:
Toys “R” Us, the future of innovation! The ad does feel a little bit wonky — and that’s even after the company admitted that it required significant post-production — but if you’re not paying too much attention, it looks rather realistic.
Kutcher’s comments and the Geoffrey the Giraffe ad sparked the same debates about the power and peril of anyone being able to generate personalized entertainment. While the pro- and anti-AI forces took their predictable stances, within just a year this idea has gone from wild futurism to one where there are several companies working on the “one prompt” movie. None are publicly available — yet — but they do exist.
So I had to wonder, how close are we to this fantastical future Ashton describes? Especially given how far we’ve come so quickly? If everyone could tailor entertainment to their own tastes, are movies as we know them soon to be a thing of the past? Should I be looking for a new job?
To figure this out, I interacted with two of the most prominent companies working on this technology: Lightricks, which makes LTX Studio, and Fable Studios, which has a forthcoming service called Showrunner. These companies, to be clear, are hoping to one day rival Hollywood, and even in their early phases we can start to see what’s to come.
Undeniably there are innumerable legal and other challenges to personalized, AI-generated content becoming mainstream. I nod to a couple of those issues in the course of this piece, but we have to start with the technological feasibility of this vision. If the tech doesn’t work, then the rest is moot.
In this article, I will discuss:
What “one prompt” generative AI video can do today
The different approaches that these two companies are taking to personalized content
How soon we can expect to be able to make full-length films
The major hurdles to seamless, fully personalized content — and whether they can be resolved
What personalized content means for the industry
 | A guest post by
|