Eli Collins, a vice president of product management at Google DeepMind, first demoed generative AI video tools for the company’s board of directors back in 2022. Despite the model’s slow speed, pricey cost to operate, and sometimes off-kilter outputs, he says it was an eye-opening moment for them to see fresh video clips generated from a random prompt.
Now, just a few years later, Google has announced plans for a tool inside of the YouTube app that will allow anyone to generate AI video clips, using the company’s Veo model, and directly post them as part of YouTube Shorts. “Looking forward to 2025, we’re going to let users create stand-alone video clips and shorts,” says Sarah Ali, a senior director of product management at YouTube. “They’re going to be able to generate six-second videos from an open text prompt.” Ali says the update could help creators hunting for footage to fill out a video or trying to envision something fantastical. She is adamant that the Veo AI tool is not meant to replace creativity, but augment it.
I’m pretty sure I’ve already encountered lots of these…
Sometimes they’re a normal video or sometimes they are shorts. Their titles are often a weird or incorrect description of the video. The video is either a loop of an AI voice reading out loud Reddit q&a posts or a compilation of shorter funny/shocking/etc videos. It kinda feels like the type of videos bot farms would watch just to give/get clicks.