Getting “roasted” online is a well-known trend that sees your photos or social media profiles trashed by a clever insult or two. Often times, it’s your own fault, with thick-skinned fans asking particularly talented roasters to target them. Now, there’s a new roaster on the scene that’s driving hundreds of thousands of Instagram users crazy: ChatGPT. The “Ask ChatGPT to roast your feed in a paragraph” trend is exactly what the name suggests. People ask the AI chatbot to look at their Instagram page and mock them in often very specific and mean ways.
Participating is easy. You need an Instagram page and the ChatGPT app. For some reason, it doesn’t work on the web portal. Take a screenshot of your Instagram page and send it to ChatGPT with the request: “Roast my Instagram feed in one paragraph.” Then try not to cry too much. For example, the following screenshot from the TechRadar Instagram resulted in the following roast from ChatGPT:
“Your Instagram feed looks like it was curated by a tech critic who was stuck at a gadget expo for too long—complete with so many smartphone closeups that your screen will crack from boredom. From AI tricks on the latest Galaxy to what feels like the 1,000th shot of a Google Pixel, it’s like a never-ending loop of “look what I have in my hand today!” Even your “what to watch this weekend” posts seem to only take a break between unboxings. If “yet another shiny new thing” were a feed, this would be it.”
AI trendsetter
It’s the kind of simple and straightforward trend that’s widely shared on Instagram Stories, so it’s easy to see why it became so successful. And while this may not be the first time AI tools have helped guide trends and fads on social media, it’s notable that it involves text. Usually, AI-powered trends are based on image filters or generated photos of users in different settings or with a common look or style. The ChatGPT Roasting Challenge relies entirely on the chatbot’s ability to analyze images and think up humorous ways to poke fun at the images. It’s also a remarkable feat, crossing the line between joke and (as far as anyone knows) completely inappropriate comments. The guardrails set up by OpenAI seem to be working in this case.
The (non-AI) originator of the trend remains anonymous. It seems to have started with a German account for books, but the creator has not revealed who is behind the account. While it may not seem like a big deal, this trend could represent a major shift in the way we deal with AI and online content in general. Not everyone will agree with it, not even ChatGPT. When the AI was asked to critique the idea of the trend, it replied:
“Asking ChatGPT to critique an Instagram account is like hiring a robot to critique your fashion choices – it’s entertaining, but also a little desperate. It’s the digital equivalent of fishing for a ‘wild’ comment from the least wild entity out there. Let’s face it: you’re basically asking a code-based algorithm to take a break from being your personal Google and instead pretend to be your snarky friend who’s had too much to drink. The whole exercise feels like getting Alexa to throw a dig – it’s more entertaining in theory than in practice, and you’re probably just doing it to avoid the sting of a real human roast.”
