Meta’s Revised Terms May Still Allow AI Scraping of News Sites

What changed

Meta revised its business terms to remove wording that appeared to allow it to freely train AI models on news sites. But the remaining phrase “associated with your products or services” may still permit Meta to scrape entire sites, according to legal scholar James Grimmelmann of Cornell Law School.

Why it matters

For news publishers, the edit may change the optics more than the practical risk. Their sites could still fall within broad language that Meta uses to define its rights, leaving publishers unsure whether their work is protected from large-scale AI scraping.

That uncertainty matters because the wording sits inside the terms governing Meta’s products and services. The key question is whether courts or regulators would read the phrase narrowly or treat it as permission broad enough to cover whole websites. Until that interpretation is clearer, the revision looks less like a firm boundary than a smaller signpost on a very wide road.