Meta has quietly changed the rules it writes for itself. The company has dropped specific language from its commercial policy that previously allowed it to scrape text, images and content from news sites to train its artificial intelligence models, according to reporting summarized from The Verge.
The change arrives following heavy pushback from publishers concerned over copyright and intellectual-property exploitation, as news organizations argue that AI training has used their reporting without payment or permission while AI products increasingly answer questions that once sent readers to news sites.
For everyday creators and publishers, the significance is practical: a major platform removing scraping permissions from its own policy is a shift in how tech giants handle copyright when building AI models, and it strengthens the hand of outlets negotiating licensing deals or blocking crawlers.
Meta has not framed the edit as a capitulation, and a policy change alone does not settle the lawsuits and licensing fights running across the industry. But in a week when publishers are also blocking AI agents at the door, the direction of travel — toward permission, payment or both — is getting harder to ignore.
