Meta drops AI data clause after backlash over web scraping

Meta has removed a disputed clause from its commercial terms after concerns that it could use website content to train AI models. Legal experts raised questions about scraping, copyright and consent. Meta says its web crawling practices have not changed.

What sparked the Meta backlash?

As reported by Firstpost, the issue began after Meta updated its commercial terms. The wording said the company could use content from domains, web pages and URLs linked to business accounts. It could use that content to build, improve, research and train its products. The wording drew attention because it seemed to give Meta wide rights over website content. Legal experts told Nieman Journalism Lab that it could let Meta collect material from publishers with business accounts and use it for AI training.

Website owners, writers and creators also worried about how their work might be used. A page may be public, but that does not mean its owner has agreed to let AI firms use it for training. Experts also raised questions about robots.txt. This file tells automated crawlers which parts of a site they may access. The clause was removed on October 17. Its removal followed public and expert scrutiny of the terms.

Meta says its practices have not changed

Meta said the updated terms did not change how it crawls the web. It said the terms it brought together applied to ads and business products. The company said it had clarified that point. Meta said the terms did not give it new rights over third-party content. Website owners could still use robots.txt to show their preferences.

The company said the clause was misunderstood. It said the change made its commercial terms clearer, rather than starting a new way to collect online content. Meta already uses some public posts shared by adult users on its platforms to improve AI systems. Other firms also use public material. But public content remains a concern. Publishers and creators want to know if their work is used, if they can refuse, and if they should be paid.

The wider debate over AI data

AI companies need large amounts of data to build and improve their models. This has led to legal fights over books, news, images and other online work. Meta won a major copyright case in 2025. But that ruling did not settle every question about AI training. Other cases involving Meta and AI firms continue. Meta’s decision removes one disputed clause, but the wider debate remains. Public access to a webpage is not the same as permission to use its content for AI training.

Meta says its crawling practices remain the same. Publishers and creators still want clear rules on consent, control and payment. The issue will likely remain as AI tools grow. The change shows how closely companies are watched when they update rules about data and online content. Wording that a company calls a clarification can still raise questions about other people’s rights. That leaves key questions open.

Author’s Opinion

Meta’s decision brings some clarity, but the bigger issue remains. People should know how their work is used online. Clear rules can help both AI firms and creators. Public content should not mean that every use is automatically accepted.

ALSO READ: Apple and Meta end years of rivalry with major AI deal.