Despite criticizing Microsoft earlier this year for using Twitter data to train artificial intelligence, Elon Musk announced today that X (formerly Twitter) would soon begin using user posts to do just that.
According to a new privacy policy that takes effect at the end of September, X could use information collected and publicly available information to “help train our machine learning or artificial intelligence models for the purposes outlined in this policy.”
Also: The best AI chatbots: ChatGPT and alternatives
That specific verbiage was absent from the last version of the terms. It’s not clear what AI model X is referring to, but Musk did launch his own AI company just a few months ago.
Musk responded on X saying that his company would use “Just public data, not DMs or anything private.”
This apparent change of heart comes only a few days after Meta announced user data from its applications (including Facebook, Instagram, and X rival Threads) would be used to train AI for its upcoming chatbot, which we could see as early as September.
While TikTok and Snapchat have both announced chatbots, neither one has mentioned taking user posts to train AI. Snapchat’s AI chatbot, My AI, does use its own conversations for training, but not general posts.
Also: How to use Claude AI (and how it’s different from ChatGPT)
YouTube uses AI to analyze and recommend videos, and other companies have used YouTube videos to train artificial intelligence, but YouTube itself hasn’t made any statements about using videos to train AI.
Why do these social media sites need user data? Artificial intelligence programs, like ChatGPT, require an immense amount of data to train properly. And the more “human” that data is, the better.
If you’re so inclined, Meta does offer a form that lets you opt out of your data being used by third parties for generative AI training. It’s not exactly clear if choosing that stops Meta’s own companies from using your data, as that wouldn’t be a third party.
