Social Network X has changed its developer agreement to prevent third parties from training large language models using the platform's content.
In an update on Wednesday, the company added rows based on “reverse engineering and other restrictions.” This is a subsection of usage restrictions. […] “Fine tweak or train your foundation or frontier model using the X API or X content.”
The change comes after Elon Musk's AI company Xai acquired X in March. Naturally, Xai doesn't want to provide competitors with free access to social platform data without a sales agreement.
In 2023, X changed its privacy policy to train AI models using public data on its site. Last October, we made further changes to allow third parties to train the model.
Reddit has also introduced safeguards against AI Crawlers, and last month browser companies added similar terms to their AI-focused browser DIA terms of service.