When Open AI launched Chat GPT it created an almost instant panic around the huge potential and pitfalls of artificial intelligence technology that can generate text becoming publicly available.
Now the company wants to do the same thing with video.
Open AI has launched text-to-video tool Sora for testing, with the intention of one day making it available to the public, just like ChatGPT.
Sora can generate videos up to a minute long while maintaining visual quality and adherence to the user’s prompt,” Open AI said on its website.
The technology is not publicly available yet but it is open to experts known as “red teamers” for testing.
So what are the potential impacts of generative AI video hitting the market?
We asked cybersecurity expert and AUCloud CEO Peter Maloney to talk us through what he believes is most likely to happen.
“I think the whole universe has only started to scratch the surface as to what are the uses of AI,” Maloney told 9news.com.au.
“You think about all the things we do in our daily lives and the way that AI can influence and impact and make things more efficient… I think it’s only just started.”
Not all AI is bad AI, he said.
Maloney said a huge benefit of the technology was its ability to quickly generate content, of which video is now the number one form on social media.
“If you’re a start-up business and you haven’t really got a marketing budget, you can’t afford to pay tens of thousands of dollars to a creative agency, this kind of does it for you… and it does it really, really quickly.”
Even though generative AI video could help businesses with their advertising and marketing, it could also hurt creative advertising industries that create video content.
He said although some basic use of Sora among friends and family would be harmless, there were potential dangers.
“The other aspect of artificial intelligence, which is what most of the regulatory authorities are really looking at, is ‘how do you do ethically?'”
Maloney said a video could be generated to give the perception of a person and sent to a vulnerable person, in the form of a romance scam.
“They believe the content that they see,” he warned.
He said it could also be used for political influence.
”In this upcoming US election with this technology, you could be watching live TV and watching Donald Trump live in Las Vegas at the same time he’s live in New York.
“Us, as the consumers, are going to have to try and translate which of those images is real.
“If you follow a news story, video imagery could be created within seconds of that news story and republished online to give a different perception of that news story.”
Maloney said generative text-to-video was different to deepfakes in the way it is manufactured.
“Those that are doing deepfake videos are really using different tooling, tooling that most consumers ordinarily wouldn’t be accessing themselves in their day-to-day lives.
“A tool like Sora is really a democratised tool that’s available for all consumers to use.”
