If you’ve watched videos on the internet any time over the past year or two, you’ve encountered AI-generated content (even if you didn’t realize it). While some of it is convincing, a lot of it is obvious. Like, I’m guessing no one thought those dramatic narrative videos of people made of fruit were painstakingly animated by hand. This type of artificially generated content has been lovingly (or not so lovingly) labeled “AI slop,” which sums it up quite well: It’s meaningless drivel, made with little effort or investment by some AI generator, that’s raking up real money by going viral on social media.
While all platforms that showcase short-form videos are inevitably already filled with AI slop, YouTube has been hit particularly hard. YouTube Shorts can be borderline unwatchable, depending on your algorithm. (Try watching in an incognito window to see what the platform serves up to a blank slate; it’s wild out there.) What’s worse, YouTube’s algorithm shows these videos to kids. Short-form videos are already bad enough for kids in so many ways, but AI slop takes it to another level.
“Did this feel like AI slop?”
There’s some good news on this front, at least: As reported by Dexerto, Starting this month, it appears YouTube is now asking users for their help identifying this AI-generated content. In fact, the company isn’t mincing words or beating around the bush. If you’re called to action, you’ll receive a pop-up that literally reads: “Did this feel like AI slop?” According to a screenshot from this Redditor, you’ll have the options of responding with any of the following: “Not at all,” “Slightly,” “Moderately,” “Very much,” or “Extremely.”
Why is YouTube asking for your help identifying AI slop?
That might sound like a good thing, and I certainly hope it is. In an ideal world, YouTube would take the results from these findings and remove the videos that are egregious. It tracks with some of the company’s past actions: YouTube has even taken down popular AI channels in the name of “reducing the spread of low quality, repetitive content.”
What do you think so far?
But Dexerto highlights a different theory, one that is less hopeful. Perhaps the reason YouTube is giving you so many choices in its popup is to understand more precisely how viewers interpret these AI videos. YouTube will then be able to tell which videos are obviously AI slop, versus videos that exist in the uncanny valley. Using that data, they’ll be able to train their own AI video models to generate content that doesn’t come across as slop to most viewers. If you’re scrolling on Shorts and the videos you see feel real, how likely are you to question whether or not they’re legit?
I hope that’s not the case. While my faith in tech companies remains low, I was impressed by YouTube’s recent stance against AI slop, and these flagging pop-ups seem to be another step in the right direction. With any luck, YouTube will continue to push higher-quality, human-made content, and crush low-quality AI slop.
