Meta removes language that may have allowed it to scrape news sites
The fact that this language existed in the first place is the real story. Somewhere in Meta's legal department, someone drafted terms that gave the company wide latitude to vacuum up web content, and it took a public outcry to get it walked back. That tells me the default instinct at these platforms is still to claim as much as possible and retreat only when caught.
I'd want to know what else is still sitting in those terms that hasn't triggered a backlash yet. This revision was narrow, and Meta's overall position on using public data for AI training hasn't budged. Publishers have been pushing for licensing deals and clearer boundaries, but the legal framework here remains murky. Removing a sentence is easy. Establishing that platforms need consent before training on someone's content is a much harder problem, and that's the fight that actually matters going forward.