How to know if a video is generated or genuine

SAN DIEGO (KGTV) - Users on X, Facebook, or Instagram are facing a whole new wave of misinformation: videos generated by AI, showing drone strikes, missile strikes, and violence. Typically paired with polarizing or false information in the captions, it can be tough to determine which videos are real and which are fake.

Here are some ways of combating misinformation and AI-generated “slop.”

Lateral Fact Checking

When you come across a shocking video online, it’s a good idea to check that video against trusted news sources, says Dr. Darby Vickers, a professor with the University of San Diego who specializes in the ethics of AI.

She says the first reflex of users should be to fact-check the video, rather than rely on built-in tools like Grok or other AI tools, because it can lead to errors in fact-checking or redundancy that makes the problem of misinformation worse.

“If you create an AI Tool to be able to detect AI material, it also creates the possibility for that same kind of adversarial learning,” Vickers said. “It creates this horrible whack-a-mole problem.”

Chrome Extensions

If fact-checking fails, there are browser extensions that can be used to indicate whether the video is AI or not. Hive is user-friendly, but there are other options, like InVID, which essentially does the same thing.

The Coalition for Content Provenance and Authority, or C2PA, has developed standards for AI-generated content, including a unique metadata “signature” that can be used to determine if the video is generated or genuine.

Features on X

X has cracked down on AI-generated or out-of-context video, mainly through a feature called “community notes.” Simply click on the top three dots of a post to request context or clarity on whether the video is AI-generated or taken out of context.

Nikita Bier, the head of product of X, also said the platform would be demonetizing any accounts that had been posting AI generated video of violence.