Meta Buys Itself Time, Not a Fix: Publishers Have Until November 20 to Force a Rewrite
Meta pushed its crawling terms back to November 20 without changing the disputed wording, giving publishers 25 days to object, price the rights and start logging.
Meta Delays Its Crawling Terms to Nov. 20, but the Wording That Worries Publishers Is Unchanged
Meta moved the start date of its new business terms from October 30 to November 20, telling Reuters the extra time lets it explain “what’s changing and what’s not.” The disputed wording is reportedly unchanged. The terms let Meta send automated crawlers to websites linked to business accounts and reuse the content for advertising and business products. Businesses accept them by continuing to use Meta’s tools.
Pressure is building on several fronts. The News/Media Alliance, which represents more than 2,000 U.S. publishers, demanded withdrawal and asked for a public answer by today. European publishers asked EU and U.K. regulators to suspend the terms, and the U.K. competition authority has asked Meta to explain them. Meta dropped the word “train” on October 16; publishers say the wording still allows AI training.
Meta did not retreat; it rescheduled, and the real contest is now over who gets to set the default and who must prove they never agreed to it.
- A default, not a deal.The terms bind by continued use, so a publisher with a business account has already “agreed” unless it leaves. There is no negotiation and no opt-out short of abandoning Meta’s business tools.
- A regulator without a hammer.The U.K. authority has engaged but cannot formally block the change because Meta has no strategic market status, so the weight shifts to the European Commission and the Digital Markets Act complaints. The leverage that matters sits in Brussels.
- A weaker licensing floor.If a platform can claim crawl rights through its own terms, every paid license it signs elsewhere is priced against a free alternative. I learned this early in media corporate finance at an investment bank: deal terms are set by whoever controls the forecast.
Use the 25 days before November 20 to put dated objections on record, price what Meta’s crawlers would take, and decide which accounts you can afford to keep, because every license you sign elsewhere gets harder to defend once these terms are live.
- LegalIndustry PositioningFile a written objection with Meta and the relevant competition authority before November 20, citing the lack of an opt-out and the unchanged wording, and coordinate with the trade body leading the response.
- BuildDetection & TelemetryStart logging Meta-ExternalAgent requests now and compare them with your robots.txt rules, so you hold a dated baseline of what is fetched before and after the terms take effect.
- CommercialEconomic ModelEstimate what the pages linked to your Meta business accounts would fetch under a license, so you can put a number on the table if Meta ever offers a deal.
Publishers on Both Sides of the Atlantic Press for a Stealth-Bot Ban as Vendors Say Most Bot Traffic Cannot Be Verified
U.S. and U.K. publishers are pushing for laws that would force AI crawlers to identify themselves and state their purpose, after voluntary controls such as robots.txt proved too weak. The vendor figures show the scale: scraping made up 70.9% of malicious bot traffic in DataDome’s September report, and Centinel found 84% of requests claiming to be GPTBot were fake. The Financial Times reported one account trying to reach about a million articles in eight hours from roughly 700,000 IP addresses. No bill has passed in either country, and most of the figures come from vendors.
The Operator’s Take: A law that makes crawlers declare themselves gives enforcement something to measure, but only the publisher that logs requests can show the gap between what a bot claims and what it does.
Build · Detection & Telemetry Log every AI crawler request with its claimed identity, source network and pages fetched, and test a sample against published verification lists so you can document spoofing before any bill passes.
The Daily Mail Tests AI-Visibility Work for Brands Ahead of a Planned Product
The Daily Mail is running 12 to 15 generative engine optimization (GEO) tests for brands across the U.S., U.K. and Australia and plans to make AI visibility an official product early next year. The method tracks category prompts, the sources AI platforms cite, and gaps branded content can fill. In an early test for a personal finance brand, visibility across target prompts rose from about 50% to nearly 75%. Future reports more than 30 GEO clients. The Daily Mail says the approach does not suit topics where government or medical sites dominate citations, and attribution remains a gap.
The Operator’s Take: Publisher authority is becoming a product that brands buy, but the sale only holds if the publisher can prove which citations its own work earned.
Commercial · Economic Model Scope a pilot that bundles citation tracking with branded content for two advertiser categories where your reporting already earns AI citations, and price it on measured visibility gain rather than a flat package.
CNN International Is Building to Sell Ads Through AI Agents, Targeting First Revenue in Q1 2027
CNN International is building in-house infrastructure to sell advertising through AI agents, with its first launch targeted for the first quarter of 2027 and performance campaigns as the early use case. It says agent-led deals could be 5% of its ad business and eventually reach 50%. Two competing standards overlap on 13 functions, and there is no agreed method to stop agents from gaming the system. An executive told Digiday’s Media Strategies event in London that the pieces are being built in silos that must be connected afterward.
The Operator’s Take: Seller-side agents automate the deal, not the delivery, so the publisher that can verify what an agent actually bought will hold the stronger hand.
Build · Economic Model Map which of your ad deals could be agent-negotiated first, set a floor price and a delivery-verification rule for each, and test them against both competing standards before committing to one.