AI-Generated Videos increasingly Deceive Online, Oklahoma Experts โWarn
OKLAHOMA CITY – As convincingly realistic AI-generated videos flood the internet, experts in Oklahomaโข are โคcautioning the public about the growing difficulty in distinguishing fact from fabrication. A recent viral video of โคbunnies appearing to jump on โขa trampoline, garnering millions of views โand interactions, was โrevealedโ toโ be entirely AI-created, highlighting the deceptive potential of the technology.
The proliferation of these videosโ poses a important challenge toโฃ online trust, โฃas even seemingly authentic footage can be easily manipulated.”Previously, videoโ was almost a โtruth, but now we just can’t rely on it,” explained dr. Johnโ Hassell, a professorโ at the University of Oklahoma’s Polytechnic Institute. Experts say verifying facts now requires a more โcritical approach than simply believing what โคone sees.
Dr.John Rhee, Associate Professor at the University of Central Oklahoma, suggests scrutinizing videos for subtle inconsistencies. “If we zoom in and slow down the โขvideo and voice,we mayโข distinguish subtle differences. โThere might beโ a slight mismatch between the lips and the voice, and than lighting – there might be some anomalies,” he โsaid.However, both professors acknowledge that increasingly sophisticated AI is makingโ these telltale signs harder to detect.
“You can’t reliably tell by looking,” Dr.Hassell admitted, emphasizing the need to shift focus from the โฃcontent itself to its surrounding context. He advises verifying the source โof โคthe video – “Who posted it first?โฃ Is it beingโ confirmed โฃby other outlets? Does the time andโ place fit?” – andโฃ treating shocking or novel clips with skepticism, akin โคto an anonymous tip.
“Bottom line is: context beats content,” Dr. Hassell stated.
While the technologyโฃ for creating realistic AI โขvideos continuesโค to advance – withโข new developments occurring daily – researchers are alsoโ working on tools to identify โขAI-generated content. Dr. Hassell noted that โฃapproximately 60%โค of โpeople have yet to experiment withโ generative โฃAI, encouraging themโ to familiarizeโ themselves with its capabilities. “It’s not scary,but youโฃ really need toโ no what’s happening,” โhe said.