Meta removes terms that suggested it could scrape sites for AI training
19 October 2026 · filed under d4d8a21bd997
Meta has removed language from its terms of service that had suggested the company could scrape third-party websites for AI training purposes, according to TechCrunch AI. The outlet reported that a Meta spokesperson said the revised terms do not reflect any change in how Meta crawls the web.
The removal of the terms, as reported, follows scrutiny of wording that had raised questions among content creators and rights holders about the scope of Meta’s data collection practices for training its artificial intelligence systems. TechCrunch AI’s report attributes to Meta the position that the company’s underlying crawling behavior remains unchanged, even as the specific contractual language describing that behavior has been altered.
The episode sits within a broader and ongoing set of tensions between AI developers, who require large volumes of data to train their systems, and the people and organizations who create and hold rights to content circulating on the web. Questions about what companies are permitted to collect, and under what terms, have become a recurring point of friction as AI firms expand their training pipelines.
Readers are directed to TechCrunch AI’s original report, published October 17, for the full account of the terms change and the company’s statement.