Meta’s proposed business terms were initially due to take effect on October 30, but publishers raised alarms that the language could allow the company to train AI systems on material from news sites using Facebook and Instagram.¹

Meta subsequently revised the wording, removing an explicit reference to “train” its products on scraped content, and pushed implementation to November 20. But legal experts told Nieman Lab that the rewrite did not definitively rule out AI training, leaving the core concern intact.¹

As the deadline approached, Meta communications chief Andy Stone signalled another retreat. “We’ve heard publishers’ feedback on our proposed ads and business terms update,” he said, adding that consultation could run beyond November 20 and that the company would delay implementation if necessary. Meta says its existing terms will remain in force in the meantime.¹

That assurance has not satisfied Washington. Sen. Marsha Blackburn, a Tennessee Republican, said the revised terms showed Meta was “attempting to exploit American creators and businesses.”¹ Her criticism put a conservative voice alongside publishers and the News/Media Alliance, which have pressed Meta over the potential use of their work.

Sen. Adam Schiff, a California Democrat, has taken the same basic view while demanding a clearer remedy: Meta should “compensate creators and publishers for their work.” In his assessment, “a delay isn’t enough.”²

The dispute has therefore moved beyond a calendar question. Meta portrays an extended pause as evidence it is listening; publishers and bipartisan Senate critics want the company to remove the ambiguity — and address who benefits when news content feeds AI.