SBS News reporter Shivé Prema took part in a separate and related creative experiment just last week. We tested all the above with OpenAI’s newest AI wizard, Sora 2. This AI tool was responsible for the hyper-realistic video of her standing in a news studio. It highlights the cool new technological ability to create hyper-realistic content. This development has spurred discussions not only about the use of AI technology in journalism but about media ethics and regulation.
Sora 2, OpenAI’s next-generation video model, can create videos from written prompts and generate clips that last up to two minutes in full HD or higher. It uses a person’s digital footprint to create composite images and videos of someone. In Prema’s case, the AI application scanned her facial features to deliver a clip that closely resembled her appearance, allowing her to ‘speak’ in the generated media.
The technology is purposely built with limitations to prevent abuse. Sora 2 only allows users to create videos of themselves, historical figures, or public figures who have given explicit consent for their likeness to be used. This measure is a direct answer to increasing worries about AI-produced deepfakes of celebrities and other characters that are protected by copyright.
When juxtaposed with an actual video of Prema, the AI video was incredibly life-like. Experts are speaking out about the dangers posed by the potential misuse of these technologies. Artificial intelligence researcher Toby Walsh wants to make sure we focus on the energy consumption. He’s worried that misleading videos will be created, spreading dangerous misinformation further.
“It’s going to consume a huge amount of energy, and I’m actually very worried that it’s going to be used for a lot of mischief, that people are going to make fake videos, and maybe we’re going to believe them,” – Toby Walsh
As there is no dedicated legislation regulating artificial intelligence in Australia at the moment, it uses existing laws to regulate these technologies. This absence of clear direction has led to contentious public discussion. Some contend that Australia is lagging behind other nations with its handling of the development and implementation AI technology.
Experts like Seyedali Mirjalili highlight the rapid evolution of AI technologies, suggesting that Sora 2’s capabilities may be just the beginning. We spoke with him about the bold vision informing Sora 2. He explained that it’s sort of like “TikTok on steroids” in terms of the amazing ability it has to produce huge quantities of content.
“It’s like TikTok on steroids, which you can generate AI content… I think they want to create a whole social media platform which will obviously be a lot bigger than what we have already,” – Seyedali Mirjalili
Whatever excitement the community has about Sora 2 is tempered with concern over what it will mean. Walsh acknowledged that this new technology has plenty of amusement value. The risks presented by deepfake imagery and political disinformation are extremely dangerous.
Mirjalili went on to address how simply people can weaponize the likenesses of public figures. “You can now upload a video of someone just half a minute, as short as a half a minute, and then impersonate them,” he stated. The technical capacity to produce inauthentic content at scale undoubtedly complicates broader ethical implications concerning identity and consent in an increasingly online environment.
“The problem is that this is Sora 2. Imagine what Sora 20 would look like,” – Seyedali Mirjalili
Despite these worries, Mirjalili still believes that Australia is not at a complete backfoot when it comes to regulating AI technologies. He believes the country is making strides compared to others and noted that there is still time for lawmakers to establish regulatory frameworks before these technologies become ubiquitous.
Shivé Prema’s experience with Sora 2 is a testament to the thrilling promise of AI-generated content. It forces us to look squarely at the challenges and responsibilities we need to confront as we welcome these new innovations. As the technology evolves further, so too must the discussions surrounding its ethical use, regulation, and the societal impacts it may bring.
