Sora OpenAIs latest video tool worries YouTubers and media

“Sora”: OpenAI’s latest video tool worries YouTubers and media

The introduction of a new artificial intelligence tool that can easily create one-minute videos by OpenAI, the Californian start-up that democratized AI with ChatGPT, raises questions, even concerns, across entire areas of artistic creation of the media.

• Also read: [À VOIR] The creators of ChatGPT introduce Sora, an ultra-realistic video generator

This software, called “Sora”, can produce “complex scenes with multiple characters, specific types of movements and precise details” from a simple line of text, assures the company, which, however, pointed out some of its current limitations pointed to confusion between left and right.

Overview of the initial reactions within the industries that could be affected by this tool:

Video creation

Among the sequences revealed by OpenAI on Thursday evening is that of an imaginary creature worthy of a 3D animated film near a candle. Another image, this time photorealistic, shows a man walking through a huge room.

It is through these two examples that “Sora’s” ability to shake up the field of video creation emerges.

“We followed the development of the imaging industry from a very early stage. It sparked a lot of internal debate, a lot of sometimes somewhat knee-jerk reactions from creative people. There were those who felt it was an unstoppable wave advancing at incredible speed and others who did not want to see it,” Thomas Bellenger, one of the founders of Cutback Productions in 2007, told AFP.

This company, which has notably contributed to the Stromae and Justice tours, specializes in the use of images and “motion design” (animation of an image, editor's note) on a large scale, for example for a concert or an immersive exhibition .

“Nobody has tested OpenAI’s new product yet. (…) What is certain is that no one expected such a technological gap in a few weeks. “This is unprecedented,” emphasizes Mr. Bellanger, but assures that in the future “we will find ways to create in other ways.”

Video game

The video game sector is also likely to be disrupted by this technological advancement and is currently divided.

French giant Ubisoft welcomes a “leap forward”. “We have been exploring this potential for a long time and as creators of worlds and stories, we see that many future opportunities open up for our players and teams to express their imagination and creativity ever more faithfully,” the group told AFP.

“My position is that its use is currently relatively unscrupulous. (…) I have no intention of replacing my fellow artists with these tools,” says Alain Puget, head of the Nantes studio Alkemi, emphasizing that “AIs only reproduce things made by humans.”

However, he insists that this “visually very impressive” tool could also be used by smaller development studios to create images with more professional rendering.

If video sequences are only a limited part of a video game intended to advance the plot, Mr. Puget still assumes that in the long term, “tools like Sora or other generative AI that produce videos will eventually find their way and replace the way will do things the way we do.”

media

Basile Simon, a former journalist and current researcher at the American University of Stanford, describes “a leap forward in the past year that is frightening.”

He fears the possible use of this tool during election periods and fears that the public “no longer knows what we can believe”.

Julien Pain, host of the fact-checking show “Vrai ou Faux” on Franceinfo channel, said he was “worried”. “Until now, it was quite easy to detect false images, for example by noticing faces in the background that repeated themselves quite often. What this new software does feels like a new level. We don’t have a magic solution,” he explains.

“There is an idea among fact-checkers to mandatory 'watermark' videos (editor's note) to indicate that they are AI. OpenAI can respect this. But (what about) Chinese or Russian competitors tomorrow?” he asks.

Advertising

As for the Fred & Farid agency, which has already worked with the brands Longchamp or Budweiser and where an AI studio was opened at the beginning of January, we assume that “80% of the brands' content is generated by artificial intelligence.” ” “The creative genius is once again the focus, production will no longer be an issue,” we were also happy.

Stéphanie Laporte, director and founder of the advertising and influencer agency OTTA, sees this as an element capable of “forcibly changing the industry” and also expects “breaks on the production side” in companies that use these new tools If you want to have a low or medium budget.

In her opinion, a possible exception is the luxury segment, “very sensitive to authenticity” and whose brands “will probably use AI sparingly.”