Is that an AI video? Here’s how newsrooms can spot them

Remember when we laughed at AI-generated images with six fingers? That was so last year. Today, video generators like OpenAI’s new Sora 2 model create clips so realistic they’re fooling millions daily, including newsrooms. A concerned community member or a prankster? The difference matters, and spotting it requires new verification protocols.

Columbia Engineering researchers recently developed DIVID, a detection tool achieving 93.7% accuracy identifying AI videos from platforms like Sora, Pika and Runway Gen-2. But journalists can’t wait for technology to catch up. The implications for newsroom operations are immediate and significant. 

Ways newsrooms can identify AI-generated video 

1. Check the video quality, especially if it’s bad.

Click here to read more.