Elon Musk’s social network X just quietly gave itself permission to use everyone’s posts on the former Twitter to train Grok, an AI model built and released by X.ai, another Elon Musk company. It was first discovered by a Twitter user.
“Twitter just activated a setting by default for everyone that gives them the right to use your data to train grok,” X user @EasyBakedOven tweeted. “They never announced it.”
At first glance, that’s an aggressive and acquisitive move to boost Musk’s AI capabilities. On the other hand, given the public and generally available nature of Twitter historically, tweets have likely been used by multiple other AI companies for years.
In fact, I just asked ChatGPT whether OpenAI used Twitter data when training it:
“Yes, OpenAI has used data from various sources, including publicly available data from social media platforms like Twitter, to train its models such as ChatGPT,” says ChatGPT, citing a May 2024 post by OpenAI. “OpenAI collects data from publicly accessible sources on the internet, which may include tweets and other social media content, as part of their training datasets to improve the performance and capabilities of their AI models.”
And Meta uses public posts on Facebook and Instagram to train its AI systems as well.
“Publicly shared posts from Instagram and Facebook – including photos and text – were part of the data used to train the generative AI models underlying the features we announced,” Meta director of product management Mike Clark said in September of 2023.
So, as aggressive and acquisitive as quietly activating a setting that allows all tweets/posts on Twitter, as well as engagements with Grok itself, to be used for AI training purposes might be … it’s pretty much business as usual in the AI space.
If you’d like to turn the feature off, here’s how:
- Go to X.com on the web (you currently can’t changed the setting on mobile)
- Click on More under Profile in the left menu bar
- Click on Settings and Privacy
- Click on Privacy and safety
- Scroll down to Grok and click on it
- Uncheck the box that says “Allow your posts as well as your interactions, inputs, and results with Grok to be used for training and fine-tuning”
Obviously, announcing the data use would be polite, even after the fact. At minimum, Meta and OpenAI have done that with their public statements on AI training data.
But overall — as someone who has unwittingly contributed tens of millions of words to AI training — I’m generally OK with help the world get just a little bit smarter.
