Meta delays new business terms after publisher pushback on AI scraping

Meta is slowing down the rollout of new business terms after publishers pushed back over what they saw as an open door for AI scraping. The complaint was simple: the wording appeared to let Meta pull material from news sites that use its platforms and use it to train AI models.

That pushback, first reported by Fast Company, was enough to force a pause. In a statement to The Verge, Meta spokesperson Daniel Roberts said the terms will now take effect on November 20, giving the company more time to explain “what’s changing and what’s not” to partners and stakeholders.

The original deadline was October 30. That’s a short delay, but it’s also a sign that publishers are paying close attention to the fine print when AI training and platform terms start overlapping. The fight here isn’t over whether Meta wants more data. It’s over whether the company can spell that out without setting off alarms.

Meta has not gone into detail in the statement about how the terms will be revised, only that the company is buying time to explain them better. For publishers, that leaves the core worry intact: if the language is loose, the scraping fight just moves from policy pages into the open.