(NewsNation) — Viral AI-generated videos are portraying some major politicians as armed criminals, with both former President Donald Trump and Vice President Kamala Harris appearing to commit robbery before being arrested.
Yonatan Dor, cofounder of the visual studio behind these videos, joined “NewsNation Prime” to break down AI’s potential and explain why the videos were generated in the first place.
The Dor Brothers videos
Dor is one of the people behind The Dor Brothers, a large name in the AI-generated visuals industry. Its recent videos of gun-toting politicians are just the latest in a long line of productions.
The studio has generated over 150 clips and garnered tens of millions of views since it began working with AI in 2022, according to its site.
“I just thought it would be funny to kind of show those figures that we all know very well doing kind of more basic things that you might see for somebody not in their status,” Dor told NewsNation. “And kind of imagine them in a different light, and just show that all humans are humans in the end of the day.”
But these aren’t just basic videos — others, like Elon Musk, President Joe Biden and former President Barack Obama, are also shown wielding weapons in the phony footage. Meta’s Mark Zuckerberg even has an iguana tagging along on his heist.
Dor told NewsNation that these deepfakes should be made without harsh restrictions, like on X’s Grok feature, rather than “immediately censoring everything.”
“I think it’s better to let people test the waters and do crazy things like what I just did, and let the public react to it and think about it,” he said.
The sordid nature of the videos is something Dor took into consideration, telling NewsNation that the internet is a “place where a lot of things are not to be trusted.”
“I think I, on purpose, made this very extreme to show people, like, everything could exist right now,” he added.
‘Something like AI takeover’
Since artificial intelligence’s surge in recent years, widespread concern about its end destination has existed.
Dor acknowledges those fears, saying that it’s “not necessarily about stopping it, but it’s maybe about slowing it down and calculating better how we continue.”
“Maybe we need smart people, which are AI safety experts, to have more control than people who are kept into capitalism and money and making money out of these tools,” Dor said. “Maybe money should come second, and the future and the goodness of humanity should come first.”
Humanity, Dor clarified, could be at stake in the future — even though he knows that the worst-case scenario sounds “outlandish.”
“Imagine AI is a much smarter species than humans, and if it has control, then it’s not really a competition. It’s just, imagine you with a dog. The dog will not be able to tell you what to do anymore. We tell the dog what to do, you know, and if we become the dogs in this example, then it’s beyond our control,” explained Dor.
“And the only thing we have to hope for is that AI has good intentions for us,” he added.” I think the smartest thing to do is to stay positive because the more people feel negative and the more people say bad things, the more likely the outcome will be.”
Spotting AI videos
Dor said the trick to spotting AI-generated images is keeping an eye on the details, like missing fingers, distorted backgrounds and missing components.
But that advice might only be useful for a bit. Dor said he gives “a year, maybe less” until commonplace AI videos are indistinguishable from reality.
Until then, Dor hopes his videos, along with other AI media, are consumed by a discerning public.
“I think people’s initial reaction should be, look at what you’re seeing and before you repost it, and keep cycling it,” Dor said. “Think about it. Think if it’s actually doing something, and don’t just post something because it agrees with your thoughts, you know.”
