The future of artificial intelligence (AI) generated video is here. OpenAI, the company behind ChatGPT, released a technology called Sora earlier this year. A film generated with Sora screened at the TriBeCa Film Festival in mid-June. This is a rapidly evolving industry, and from the halls of Congress to the boardrooms of tech companies, Americans are looking for ways to innovate while ensuring they have the right guardrails in place.
“The speed of this technology is astounding,” said Claire Wardle, professor and director of Brown University's Information Futures Institute.
Sora can generate videos up to a minute long based on prompts entered by the user. The video has amazed many. “You have to realize that this technology is improving rapidly every day, every week,” Wardle says. The Sora-generated video isn't perfect yet; there are errors, like a person running backwards on a treadmill or a mixing spoon appearing out of nowhere. But Wardle expects the technology to adapt quickly.
“I don't think it'll be long before these things look really perfect, and we'll get to a point where our eyes just won't be able to make it out,” she said.
Sora's technology isn't perfect yet, but experts believe it will adapt quickly. (OpenAI)
Other companies are working on similar technology, and the competition is heating up, with a Chinese model called Kling producing particularly impressive footage.
To create a video, an AI video generator like Sora is first trained on a vast library of images, and then when it receives a prompt, it creates a composite image using all the images it has seen so far.
“It's really just copying. Finding a pattern and copying it,” Wardle explained.
The AI videos have raised ethical and legal questions: OpenAI has remained vague about where it gets its training videos, and YouTube has argued that using the company's videos violates its terms of service.
“We want to see high-quality content used as training material, but the big questions about payment, consent and other issues will come up here,” Wardle said.
Experts are also beginning to point out the potential for this technology to change the way we work.
“I don't think jobs will disappear, but they will be very different,” Wardle said. “I think we'll see new jobs like prompt engineers. How do we design prompts that will give us the right response from the AI?”
The technology also increases the potential for misinformation, especially as we approach election season. While manipulated images are nothing new, AI-generated imagery takes us into uncharted territory.
“We now know how easy it is for anyone to completely fabricate a video or an image, so when we look at it all now there's just a shred of doubt left,” Wardle said.
But the last thing Wardle wants is for people to just throw up their hands and give up.
“If we give up, turn our backs on this and don't believe in anything, that's actually one of the goals of the bad guys,” she said.
Sora isn't yet available to the public, but Open AI says it is “red teaming” the model — assessing its vulnerabilities and strengthening its protections.
As an added layer of protection, Wardle stresses the importance of sourcing your news from multiple trusted news sources and treading carefully before sharing any suspicious images or videos on social media.
“If you see something, you should look it up on a search engine first, because almost always someone will fact-check it right away and say, 'No, this is an old photo. It's real, but it's from 2016,' or 'No, this has been fact-checked. The president didn't say that last week. It was cropped.'”
As Sora and other AI technologies evolve, Wardle says it's important to remain vigilant while society learns how to navigate this new territory.
“It's really important that we continue to talk to each other, continue to check in, continue to evolve and adapt to this new technology, and that will get us through this difficult time. But we can't give up,” she said.
