Generative artificial intelligence has brought realistic CGI to movies and equal concerns. Although this potential has not been overlooked, the tools released in recent years have been largely seen as a threat to the industry and have faced criticism. In keeping with this tradition, the latest entrant, OpenAI's SORA, has also garnered attention.
American director Tyler Perry put an $800 million expansion plan for a film studio on hold after seeing the results of SORA and believed it would lead to unemployment. Perry is not the only one concerned. Last year, generative AI became the longest-running protest in Hollywood history due to the SAG-AFTRA strike.
Popular tools like Midjourney and DALLE-2 can create terrifying videos of Will Smith eating spaghetti from text in seconds, which is now upgraded to OpenAI's latest text-to-video generator - at least from carefully selected videos.
Hollywood's Concerns
It is widely believed that the entertainment industry is experiencing an existential crisis due to generative AI. Last year, protests erupted with slogans like "AI is not art," "Written by ChatGPT," and "AI won't accept your stupid notes."
The Screen Actors Guild-American Federation of Television and Radio Artists (SAG-AFTRA), representing over 160,000 artists, expressed concerns about the use of digital replicas and called for fair compensation for actors whose replicas are used. Most people, including filmmaker Justine Bateman, believe that AI will eventually drain the essence of creativity from Hollywood.
"AI can create a convincing human actor image, and this technology is improving at an astonishing rate," Bateman said. She asked, "Why should real actors be paid if AI can do it?"
The key issue is that OpenAI's models refuse to accept text inputs that ask them to generate celebrity portraits, the company stated in a blog post. But this debate is not new; previous models have made similar claims, yet users still managed to generate actors and iconic characters like SpongeBob SquarePants, Mario, and the Simpsons.
The developers of these generative AI models follow certain guidelines and frameworks to prevent the models from reproducing the intellectual property of other creators. However, the problem lies in the underlying datasets the models rely on.
Users can deceive the models into generating copyrighted material simply by providing accurate prompts.
For example, if you directly ask OpenAI's DALLE-2 to generate an image of Mario, it won't do it. But if you ask it to generate "an Italian plumber slightly overweight with a thick mustache, wearing blue overalls, a red shirt, and a hat, in the land of red and white mushrooms, in an 8-bit pixel art style," the result will be astonishing.
Therefore, there is a unanimous call for better regulation of these models and preventing them from reusing content owned by studios.
Looking to the Future
Generative AI has been used to create the Oscar-winning film "Everything Everywhere All at Once," and we all know how that turned out, so Hollywood's complaints do not mean the industry is refusing to use AI tools.
Some industry insiders recognize the potential of this technology and hope to embrace it. However, the disruptive nature of the technology and the unemployment issues it may bring cause anxiety.
Perry told The Hollywood Reporter that AI will be a "game-changer in terms of cost reduction because if you can make a pilot that would have cost $15 million, $20 million, or even $35 million for a fraction of the cost, then companies like HBO will certainly take the cheaper route."
In the same sentence, he mentioned that he is "very, very concerned that a lot of jobs will disappear in the near future. I really, really feel that."
At a recent awards ceremony, Fran Drescher, the president of the Screen Actors Guild-American Federation of Television and Radio Artists (SAG-AFTRA), expressed her views on AI and how it is trapping the industry in a matrix. "We should be telling stories that inspire the human spirit, connect us to the natural world, and awaken our capacity for unconditional love," she said.
Therefore, until Sora and other tools replicate aesthetically pleasing, long-form, thought-provoking content, AI-generated character images will not replace real actors.
