Meta removes wording that might have permitted it to scrape news websites

Meta Removes AI Training Language From Commercial Terms

Meta description: Meta updated its commercial terms page to remove language suggesting content from websites, including news sites, could be used to train AI.

Tags: Meta, AI training, commercial terms, Neiman Lab, James Grimmelmann

Featured image: abstract technology style with no real people, company logos, or text

Background
The issue centered on one sentence on Meta's commercial terms page. The removed line named three content categories: domains, web pages, and URLs. It also listed six uses: provide, develop, improve, research, train, and maintain the integrity of Meta Company Products, services, and features under those terms and other applicable supplemental terms. The wording was read as suggesting Meta could scrape content from sites, including news sites, to train its AI.

Development
The sentence became a flashpoint after backlash. Neiman Lab and James Grimmelmann, a professor of digital and information law at Cornell, were among those who interpreted it. The debate focused on how the terms could be read. Meta then updated its terms again.

Latest stage
The removed sentence no longer appears on the commercial terms page. The change narrows the visible wording, but the page remains the place where readers can inspect Meta's terms. The current stage is a correction after interpretation and backlash: Meta has deleted the specific sentence, leaving the broader terms and supplemental terms to be reviewed by readers.

What to watch
The practical signal is that one sentence in the terms page became the focus of public interpretation. Readers can now compare the current commercial terms page with earlier wording to see whether the removal changes how domains, web pages, URLs, and supplemental terms are described. The issue is now a removed sentence on a revised commercial terms page, while the broader terms still await closer reading.