Paris (AFP) – Last week, US company OpenAI unveiled a tool that can generate highly realistic video snippets from just a few lines of text, leaving content creators wondering if they are the latest professionals about to be replaced by algorithms. I started to doubt it.
Publication of: change:
2 minutes
Reactions to the tool, called Sora, have ranged from enthusiastic excitement to wariness about the future direction of the industry.
YouTuber Marques Brownlee said it was “scary” and “threatening” to watch AI do the job.
Meanwhile, Caleb Ward, one half of AI filmmaking duo Curious Refuge, told his YouTube followers he can’t wait to get his hands on the tool.
But both Ward and Brownlee agreed that this is a big moment for their industry.
“I can’t overstate how big a deal this is for the filmmaking and creative world,” Ward said. He recently made headlines for the trailer he created for a Wes Anderson-style Star Wars movie.
OpenAI, the developer of ChatGPT, said in an announcement that Sora is not yet publicly available.
The announcement did not specify any use cases, but said “a number of visual artists, designers, and filmmakers” had been selected to help with the test.
“Like an amoeba”
The company attached sample videos to its statement, including a stylish woman walking the streets of Tokyo, a cat waking its owner in bed, and a herd of woolly mammoths charging.
As is often the case with OpenAI products, the internet immediately erupted in awe and admiration.
“I was shocked by the quality,” Anis Ayari, an AI engineer and streamer known as Defend Intelligence, told AFP.
He suggested that the tool could one day be used to create completely virtual presenters.
However, there were many naysayers who felt that the video was still firmly in the uncanny valley. This means that photo-realistic images may have flaws that may cause discomfort to viewers.
“The owner’s arm looks like part of the cushion, and the cat’s paws pop out of the arm like an amoeba,” commenter Ed Zitron wrote on OpenAI’s cat video.
He wrote in his newsletter that AI video tools are too expensive and resource-intensive to be truly useful.
Also, clip styles could not be harmonized, making the tool useless for creating anything other than small pieces.
AI fatigue
Sora enters a heated market that is already joined by Google, Stability AI, and several other smaller players.
YouTube itself announced last September that it was developing tools that would allow creators to create AI-generated videos and background images.
However, the tools already available haven’t taken the world by storm.
French streamer FibreTigre said he tried an AI video tool but stopped the experiment.
He said he was concerned about the ethics of using tools trained on other artists’ work, and that ultimately the program did not work well enough.
“It’s just ugly,” he said of the AI video.
He said he could see a future where viewers would experience “huge fatigue” from AI and value things that aren’t artificial.
© 2024 AFP
