Meta's new terms of service scrape news sites; another nail in search traffic's coffin
AdExchanger · 19 Oct 2026, 06:03 CEST · Score: 10
Starting 30 October, Meta extends its scraping privileges from business accounts' own content to every website linked on its platforms, meaning business pages that share a news link effectively consent to Meta scraping that site for AI training, with no opt-out short of quitting Meta advertising entirely. Danielle Coffey, president and CEO of the News/Media Alliance, called the terms "an appalling, bad-faith policy, and a clear abuse of the company's market power", while media and IP attorney Matt Topic of Loevy & Loevy called it "a remarkable provision to bury in [the] terms of service" that "would give Meta access to material that would likely cost many millions of dollars in a licensing market". The roundup's second act is the same squeeze from the demand side: The Independent's CEO Christian Broughton said social-commerce revenue narrowly surpassed SEO commerce for the first time last month and called the old traffic-to-ads model "dead", though the roundup's own caveat is that the publisher does not disclose its revenue mix, so it is hard to say whether social commerce actually grew or SEO simply collapsed. The tail of the roundup makes the political-money angle concrete: Elon Musk has given about $190 million to the Republican Party's main Super PAC for Senate races, which has already spent $159 million on advertising for a single Texas candidate.
The crux
Meta is writing itself a legal basis to harvest publisher content for AI without a licensing fee, at the same moment the search-referral model that publishers relied on is collapsing, hitting the same revenue line from both directions.
Sell-side read
The "wait and see" approach to AI licensing now has a clear cost: Meta's terms demonstrate that the default is unpaid scraping, so a publisher either negotiates a deal or is training someone else's model for free.
Watch
The 30 October effective date, and whether legal challenges built on robots.txt supersession or market-power abuse gain traction.
Published: 19 Oct 2026, 06:03 CEST