OpenAI has announced a generative artificial intelligence tool capable of producing hyper-realistic videos, marking the San Francisco startup’s first major foray into Hollywood.
The system appears to be capable of producing movies of complicated situations with numerous characters, a variety of shot types, and generally accurate descriptions of subjects in relation to their backdrops. A presentation released on Thursday featured brief videos that the company claimed were made in minutes in response to a text prompt of only a few phrases, though several were inconsistent. It featured a movie clip of an astronaut traveling through a desolate planet and an animated scene of an expressive creature kneeling before a melting red candle with shadows floating in the backdrop.
READ MORE: The FTC Is Looking Into ChatGPT Producer OpenAI For Potential Consumer Harm
While similar AI video tools have existed, OpenAI’s approach demonstrates rapid advancement of the technology, which has the potential to undercut large segments of labor. It signals increased mainstream usage as the entertainment sector grapples with AI. A January study of 300 Hollywood leaders found that three-quarters of respondents said AI tools favored the loss, reduction, or consolidation of positions at their companies. Nearly 204,000 jobs are expected to be lost during the next three years.
READ MORE: According To The Information, OpenAI Is Planning An App Store For AI Software
According to the report, sound engineers, voice actors, and concept artists led the way in this displacement. Visual effects and other postproduction tasks were also identified as especially risky. These occupations, among others, are increasingly viewed as being replaceable by AI tools as technology advances.
“This is a clear alarm to the unions and professionals who are in crew in any capacity,” says Karla Ortiz, a concept artist who has worked on multiple Marvel novels. “This demonstrates that technology is ready to compete with us.” This is just the first step.
Reid Southen, another concept artist who has worked on The Hunger Games, Transformers: The Last Knight, and The Woman King, emphasizes that many of his contemporaries are seeing decreased demand for their work.
“I’ve heard a lot of people say they’re leaving film,” he said. “I’ve been thinking of where I can pivot to if I can’t make a living out of this anymore.”
Whereas most AI video generators are limited in the length of videos they can produce, which frequently have issues with extra limbs or demonstrating real-world physics, OpenAI’s tool, Sora, appears to be close to generating content up to a minute long that maintains visual quality and consistency while adhering to user commands. It allows you to transition between shots, such as close-ups, tracking, and aerials, as well as change shot compositions.
However, there are several drawbacks. In a video of a woman walking down a Tokyo street, her legs switch sides halfway through the footage. The lapels on her jacket also appear to shift between wide-angle and close-up.
“The current model has flaws,” OpenAI explained in a blog post announcing the tool. “It may struggle with accurately simulating the physics of a complex scene, and may not understand specific instances of cause and effect.”
According to the report, “The model may also confuse spatial details of a prompt, for example, mixing up left and right, and may struggle with precise descriptions of events that take place over time, like following a specific camera trajectory.”
Sora will not be distributed to the public while it is undergoing safety testing. It has been shared with specialists in misinformation, hateful content, and bias, as well as visual artists, designers, and filmmakers, to solicit comments on improvements.
READ MORE: OpenAI CEO Sam Altman Admits To Being “A Little Bit Scared” Of Artificial Intelligence
OpenAI also stated that it is developing tools to assist spot deceptive information generated by Sora. It intends to include metadata in videos produced by the system that will identify them as being created by AI. Furthermore, the tool will reject prompts that violate usage policies, such as those requesting severe violence, sexual content, celebrity likenesses, or the intellectual property of others.
The corporation did not reveal the materials used to train its system. It no longer discloses the source of training data, citing the need to preserve a competitive advantage over other enterprises. Several authors have sued the corporation, accusing it of utilizing their copyrighted publications, the majority of which were downloaded from shadow library sites, according to multiple complaints. If some of these suits succeed, they could force OpenAI to remove systems trained on copyrighted information.
In an indication that AI may be rapidly evolving using copyrighted content, AI picture generator Midjourney began returning virtually exact reproductions of frames from films following its most recent update in December. When prompted with the phrase “Thanos Infinity War,” Midjourney, an AI tool that transforms words into hyper-realistic images, produced an image of the purple-skinned villain in a frame that appears to be from the Marvel film or advertising materials. It could also appear to duplicate practically any animation style, creating characters from a variety of titles, including DreamWorks’ Shrek and Pixar’s Ratatouille.
“It’s going to destroy our industry,” Ortiz claims. “Everyone is pitching generative AI projects right now. We are already getting screwed.”
Commercial production may be one of the first casualties of AI video technologies, as quality is deemed less crucial than in film and television production. Commercial production in Los Angeles has largely declined since 2018, reaching a seven-year low in 2023, excepting 2020, when the pandemic shut down most productions.
Radiant and America Nu, offering to elevate your entertainment game! Movies, TV series, exclusive interviews, music, and more—download now on various devices, including iPhones, Androids, smart TVs, Apple TV, Fire Stick, and more.
