Can OpenAI’s Sora-to-video generator block jobs in Hollywood? | | Technology News


Artificial Intelligence startup OpenAI has been teasing its new AI video generator, Sora, on social media in recent weeks. Last week, it revealed that it had also given Hollywood actors and directors a first look at the technology – and a chance to try it out – before Sora’s public release.

OpenAI published a blog on March 24 titled Sora’s debut, showing the work that several production studios and directors have created using the video generator.

Some media experts believe that Sora will disrupt the film industry.

Al Jazeera spoke to an executive who works in Hollywood, who asked not to be identified because of the complexity of the story. When asked what his first reaction was when he saw Sora’s powers for the first time, he said: “My reaction to Sora was like everyone else’s – my jaw dropped. It was like seeing our killer but it was beautiful at the same time. Fascinating and terrifying at the same time. “

The tremors that caused Sora are already being felt by some in the industry.

In an interview with The Hollywood Reporter in February, actor, film producer and studio owner Tyler Perry said he would stop expanding his $800m studio in Atlanta after seeing Sora’s filmmaking skills.

He added: “So I am very worried that soon, many jobs will end. I really feel that very strongly. “

What is Sora?

Sora is OpenAI’s AI version of text-to-video. Similar to ChatGPT, you enter a quick text but instead of providing answers to questions or queries in words, Sora creates videos of up to one minute.

A video example of Sora’s abilities, released by Open AI, can be seen below:

  • Example: “Movie trailer showing a 30-year-old spaceman in a knitted woolen helmet, blue sky, salty desert, cinematic style, shot on 35mm film, vivid colors.”

Sam Altman, CEO of OpenAI, posted several examples to his X account, including the one below:

  • Note: “A recipe for homemade beekeeping by an activist grandmother in a Tuscan kitchen illuminated by video.”

Sora is perfect. If you watch the “cooking tutorial” video carefully, the spoon in the right hand disappears after “grandpa” stops mixing. Even though it’s true, the possibility of false positives still exists in some of the videos Sora makes.

This raised another question as to whether something like this would work well in the industry.

Our Hollywood director had this to say, “We’re going to come out bigger and better because people will have realized their place in a technology that’s far more powerful than we can imagine. But the desire to act, write, direct, compose, collaborate, etc. is natural by nature. people. It’s not going anywhere. So is it bad for the people in the industry? The answer is no, then yes, then no. Is it good for the industry itself? Yes.

How does Sora work?

As is ChatGPTusers write a command, a question or a prompt and the AI ​​responds – in the case of Sora, with generated videos.

To do this, Sora uses a combination of machine learning and natural language processing (NLP) to create videos. NLP is a form of artificial intelligence that understands the interaction between computers and human language. Machine learning makes Sora better over time while improving its responses through actions and feedback.

Sora uses “computer vision” to understand and interpret visuals from images or videos. Computer vision is a software program that tells Sora to “recognize” representations of real-world objects, people and environments from verbal descriptions that include visual language. For example, the words “the cat is walking” or “the waves are crashing in the sea” indicate certain conditions and behaviors. Sora needs this visual language to quickly translate the words, and then accurately display the image of an object.

Sora is able to harvest unrealized or semi-realistic footage and turn it into vivid animations that look real. Sora serves as a very powerful visual tool. It starts with large, abstract blocks of color or material and then breaks them down into smaller, more defined blocks based on your needs.

What does Sora mean in the film industry?

It is not yet known whether, if any, the works of the people who created it will be taken over by Sora. The ability of AI to repeat camera shots, lighting and characters on the fly allows directors and filmmakers to enter the unknown. However, film experts hope that this will have a major impact on the industry.

A Hollywood insider who spoke to Al Jazeera on condition of anonymity said: “I don’t see it as a threat to production without a threat to the way production is done as we know it now. We’ve seen things like this in the past, especially in production after people started switching to laptops in place in the big expensive houses. Many people were completely exhausted by the change while others were able to buy the right editor without the need for a post house.”

When asked which jobs could be replaced by AI generators, he added: “Maybe asking ‘who will replace’ is the wrong question. I think it is a system that can be replaced and replaced. In a few years, maybe the word ‘leader’ will mean the person who promotes AI, and the rest is done digitally. And if this method is accepted by the audience, makes money, and makes people feel human – then it’s over for most of us.”

What are copyright and legal issues?

Sora pulls content from pre-existing photos and videos, then recreates the video based on the user’s input. Who exactly has the video edited? Should money be paid to everyone who creates the graphics and animations and characters that Sora uses to create the final movie? These are questions that have not been fully answered.

At the root of many of the above questions is how to track the originator of each product, including the people included in the last video.

Speaking on his YouTube channel, technology lawyer Paul Haswell, explained: “If someone is just using an AI model and then unknowingly in some way looks like you, what are your rights – then your data is being misused. ? How can you prove it? that your data was used to create this?”

He added: “Suddenly you find yourself in an AI soap opera. You can be world famous, but you don’t get fame for it. You can have a catchy voice and not a deep voice but your face can be the same. For example, you can’t have a debt because you are used.” work with AI, to be standardized and transformed into a different kind.”

There are also international considerations as copyright law differs from country to country. If the video came from one country and was sent to another country, whose law applies?

On his blog, Wallace Collins, an entertainment lawyer who specializes in copyright and commercial law, warned that Sora would exacerbate all of these problems “increasingly” and could lead to riots or other types of problems. social disruption.

“AI has already undermined copyright law for creators, especially in the music industry, and has challenged the traditions of espionage and intelligence in entertainment. Without some kind of clear rules, Sora can be used by the worst people to create videos that can defame, mislead and threaten. people, or start a riot by looking at the appearance of something fictional but real. appearance.”

How will these issues be decided?

A large part of the legal debate surrounding AI output also concerns the issue of who should be considered the author of what these tools produce, as it relates to “fair use”. Fair copyright laws allow the limited use of copyrighted material or the adaptation of a copyrighted work to another work.

Currently, there are no regulations regarding the latest advances in audio and video production. However, in December last year, the New York Times wrote a federal case against OpenAI’s ChatGPT (a text-to-speech tool) and Microsoft’s Copilot for copyright infringement in the Southern District of New York (federal court in Manhattan). The Times reports that OpenAI’s ChatGPT offers users the same content that the Times previously offered.

Ian Crosby, a lawyer for the Times, said: “The defendants want to fight the Times’s large investment in its journalism by using it to create substitute material without permission or payment.” That should not be fair use.”

In February, OpenAI filed a motion to dismiss the Times lawsuit in federal court.

Two other copyright infringement lawsuits were filed in Manhattan court against OpenAI — one by The Intercept and the other a joint lawsuit by Raw Story and AlterNet — in February.


Source link

Leave a Reply

Your email address will not be published. Required fields are marked *