Viral video AI concerns of bunny jumping to trampoline: expert weight

AI Video & Visuals


The video of a rabbit hopping on a trampoline not only caused a strong online response, but also caused a strong response.

The 8-second clip has received over 25 million likes on Tiktok since it was posted on July 26th. It starts with six bunnies perched on a trampoline and escalates up and down onto the equipment.

Why did this seemingly simple video gain so much traction? You can get the answer by looking through the comments. It can be generated by AI.

Eagle Eyed viewers called for details in the comments. At the beginning of the clip, one of the rabbits in the lower left corner appears to move from two heads to one, and at the 6-second mark, one bunny appears to be completely gone.

“One of the bunnies had two heads first…” one wrote.

“I almost got me until one of the bunnies magically disappeared,” the second commenter said.

“They were six bunnies, one disappeared… it's AI,” a third user wrote.

Some people said the video was bewildered because of how realistic it was that they first found it.

“This is the first AI to win me,” one user commented on the video.

“A few years ago, I was laughing at my mother, believing in AI,” one viewer said.

“I'm not the only one who thought this was true,” another added.

The original creator of the video has not confirmed whether the video is AI or whether it has released details about how the video was created. But Cara Badiro Urquiora, a professor of computer science and engineering at the University of Notre Dame, points out “clues” that support video.

“If you look closely, you can see some bunnies appear and disappear, or they come and go from each other,” she tells Today.com. “These types of inconsistencies and unnatural behaviors are common in AI-generated videos or images.”

Researchers at Northwestern University published guidelines on how to identify AI-generated images in 2024, and the findings support allegations about video.

Research shows that unrealistic anatomical qualities, such as the two-headed rabbit, can indicate that images and videos are being generated by AI. Furthermore, paying attention to the fusion of body parts, which can also be seen in Tiktok videos, is useful when identifying AI.

The current video glitches not only refer to it being generated, but the bunny's actions are inconsistent with the real rabbit's actions.

Seth Muggle, director of the Urban Wildlife Institute at Lincoln Park Zoo, told Today.com there are some indications that the “rabbit” behavior in the video isn't like a normal bunny.

“I would be very surprised if a rabbit, a prey animal, a rabbit, an animal that must be extremely vigilant about unknown threats and dangers, wherever they live, jumps up probably a foot and a half to two feet tall,” says Muggle.

He adds that it is unlikely that multiple rabbits will find interesting by jumping on unfamiliar, bouncy surfaces.

“It doesn't match what we know about rabbit behavior,” adds Muggle. “They naturally fear new things.”

David Rand, professor of management science and brain and cognitive science at MIT, tells Today.com that he is researching the relationship between AI's misinformation and that changing videos is not a new phenomenon. However, more advanced technology makes it easier to produce.

“When you're viewing all the videos these days, it's really easy to work with them,” Rand says.

Badillo-Urquiola adds that trust and safety should be prioritized as AI-generated content becomes more realistic.

“AI-generated content continues to become better and more difficult to find,” she says. “If that's the case, this may mean that we can reach a point where we can't trust anything, or we can't make sure the source is very reliable unless we see it with our own eyes.”

Rand adds that many experts are pushing for labeling internet videos like rabbit videos as AI generation. However, Rand says that because the video is partially adapted by AI or partially adapted using non-AI methods, there are gray areas with AI or not, such as adding visual effects.

“The real question is not, 'Is this being generated or not?',” says Rand. “The real question is, “Is this misleading? Will this lead to someone to have an inaccurate understanding of the world?”





Source link

Leave a Reply

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