Forget spam. We’re drowning in slop. That’s a new term for all that AI-generated garbage that you might have noticed on social media or elsewhere across the internet. Whenever you see crappy Google AI Overview results or Facebook photos of unnatural-looking seascapes posing as nature photography, you’re encountering the wild world of AI slop.
Not all generative AI is, by definition, slop. Just the worst uses of it. But like email spam, it’s unwanted, inaccurate, deceptive, or altogether unnecessary. Some of it is explicitly profit-driven, designed to soak up ad dollars or scam people, but some of it is just the result of popular AI models very often spitting out incorrect information or unbelievable images. It fills space, fuels confusion, and makes the internet a worse place to be. It’s already making Google and Facebook less useful by filling search pages and timelines with junk.
It can be downright dangerous. For instance, mushroom enthusiasts were recently warned to avoid fungus-hunting guides from Amazon due to the proliferation of AI-generated books on their marketplace. One bad hallucination from a bot, and you could be having some pretty wild hallucinations (or worse) of your own.
But preventing slop falls pretty far down the priority list for policymakers, so it could be years before policy meets the problem – as we’ve seen with spam phone calls, email, and texts.