Meta drops the AI scraping clause from its business terms after a backlash

[FluxKnight]

Meta has backed down over its controversial new business terms. TechCrunch's Anthony Ha reports that the company has removed the clause that would have let it use content from business customers' websites to train its AI models. We discussed the original change a few days ago, after The Register reported it

The removed wording would have allowed Meta to use any content from a business account's domains for training. Legal experts quoted by Nieman Lab had called it remarkable, and publishers were angry because almost every news organisation has a Meta business account

Meta now says the terms only ever applied to its ads and business products. It also says the change does not affect how it uses web crawlers. That leaves some questions about what the original wording was meant to achieve, but the clause itself is gone

The reversal shows that pushback can work. The News/Media Alliance had called the terms outrageous and probably illegal, and the story spread quickly. Companies rarely change course this fast unless they expect serious legal or reputational trouble

It is a good outcome for businesses and publishers. Buying ads should not mean handing over your website for AI training. The wider fight over AI training data is far from over, though

Did Meta overreach or was this a drafting mistake? And will other platforms try something similar?

Meta removes language that raised scraping and AI training concerns | TechCrunchtechcrunch.com