Stop Elon Musk From Taking Your Content to Train His AI

A common refrain in the ongoing AI discussion is that “AI is theft.” Which is understandable given that AI developers frequently train their AI tools on content without the permission or knowledge of said content’s creators.
404 Media recently revealed that the developers of Runway, an AI-powered text-to-video generator, trained their tool “by scraping thousands of videos from popular YouTube creators and brands, as well as pirated films.” Needless to say, Runway’s developers didn’t get permission from the creators of those videos and films. And just last month, I wrote about Perplexity, an AI-powered “answer engine” accused of plagiarizing content.
Which brings us to Elon Musk and X (formerly Twitter). Earlier this week, @EasyBakedOven posted about a new X setting that lets users control whether or not their posts are used to train Grok, X’s AI tool. But… X had enabled the setting by default without ever notifying users of its existence. What’s more, the setting is buried in X’s interface, meaning the likelihood of stumbling across it is pretty low, and it can’t be changed via the X app but only on the X website. All of which is more than just a little sketchy.
Note: If you’re an X user and don’t want your content used to train Grok, click here to view and update the setting.
MacRumors has more info, including when X possibly added the setting. Finally, X subsequently posted about the setting, too, explaining that “all X users have the ability to control whether their public posts can be used to train Grok” and “this setting… will soon be rolled out on mobile.”
No one should be surprised that Musk wants to use their content to train his AI tool. Human-generated content remains the best way for AI tools to “learn,” and considering the amount of content posted on X every single day, it’s practically a renewable resource. (Of course, the quality of that resource is an entirely different matter.) Nor should it be surprising that it’s done in a sneaky, underhanded manner (i.e., opting you in without any notification or consent).
After all, most people aren’t too fond of their content, and the time, effort, and creativity it represents, just being taken and used without any sort of acknowledgement or compensation. That’s doubly so with regards to AI, which looks less and less like a benefit to humanity and more like a blatant cash grab that taxes the environment and numbs our aesthetic sensibilities.