“Goodbye, Meta AI” is the latest Facebook copypasta making the rounds online. The thick wall of text, pasted against a hazy orange-and-yellow gradient background, has all the hallmarks of this trend, including vague references to legal systems and one-sided declarations of personal protection. It has a nostalgic feel, reminiscent of the days when chain letters were mandatory. But unfortunately, posting an image to Facebook, Instagram, or any other social media platform doesn’t actually allow you to opt out of having your post fed to an AI model.
This certainly isn’t the first time that nonsensical copypastas have gone viral on social media sites: over a decade ago, WIRED ran a popular “copyright scam” that blanketed Facebook with “pseudo-legal jargon.” It didn’t work then, and it doesn’t work now.
The hashtag “Goodbye Meta AI” has been circulating since early September, reportedly being shared on Instagram Stories by Tom Brady and James McAvoy. The hashtag’s purported ability to protect data is blatantly dubious to savvy internet users. But the underlying desire to get personal information back from tech companies is relatable. Companies know so much about users’ lives and aspirations that it can be unsettling. And as the wave of generative AI continues, it seems like everything posted online is at risk of being scraped to train the next most powerful and powerful AI model.
Two big red flags that help you quickly spot a copypasta like this are an urgent call to action and an unclear reference to a legal situation. In this case, the image reads “ALL MEMBERS MUST POST” to keep data safe, claiming to be part of the advice of an anonymous lawyer. The 2012 version read, “Anyone reading this can copy this text and paste it on their Facebook wall.” The decade-old copypasta also contained a misspelling to a European legal contract.
“While we don’t currently have an opt-out feature, we have built an in-platform tool that allows you to remove your personal information from chats with Meta AI across all apps,” company spokesperson Emil Vásquez said in an email. Instructions are here. He also noted that European users can object to their personal information being used in AI models, though as WIRED reported last year, the objection forms provide little to no help to users.
So if accidentally copying and pasting doesn’t work, how can you prevent your published words and images from being used in Meta’s AI models, and those of other AI companies? Just stop posting online. Other than walking away and never posting again, there is currently no realistic way for an individual user to circumvent savvy scraper bots.
With that in mind, you can take steps to reduce the amount of information you expose on your social media profiles to increase your privacy a bit. It also doesn’t hurt to download your old posts for your own records and delete most of them from the internet. If you want to go even further, here’s a list of websites and apps that allow you to opt out of at least some of their AI training practices.
