Resource · For publishers

A fair
licensing collective.

One license, many publishers, one pool of money. A collective is fair when each member can see what the pool collects and check its own share. This page sets a collective on FORA beside the eight schemes running or announced today.

Updated

Follow one article

Same money. Open boxes.

A trade group, a publisher alliance or a company working for one runs the collective as its operator. Follow one article through a pool today and through a collective on FORA.

Your article is used in an AI answer

Step A pool today A collective on FORA
What you put in
The uses the scheme defines: every AI use, or a fixed list.
Your own terms, as simple or as detailed as you want.
Picked
The AI company's or the operator's own systemNo scheme publishes how it picks for an answer
The operator's searchThe operator can publish the method
Counted
The payer's own report or engineOr no count at all
A signed record of every fetchThe payer didn't write it
Split
Undisclosed rules or weightsIn seven of the eight schemes
A published formula and published weights
You see
A statement, at most
Your own recordsYou can redo the sum

You can't look inside You can check it

The money takes the same path on both sides. The left column sums up the eight schemes below. CCC and CLA haven't said how they will count AI use, and one scheme, GPT-NL, publishes its split rule.

None of the eight schemes below publishes how an AI system picks whose content to use in an answer, and in a pool split by use that pick decides who gets paid. Who decides

The formula behind step 3

Each AI company pays the greater of a share of its own revenue and a floor per fetch, up to a cap, so the pool never falls below the floor. The formula and the per-fetch floor are published; each AI company's rate and cap are in its agreement with the operator.

The pool, each periodformula and floor published

each AI company pays
min( cap, max( rate × its revenue, floor × fetches ) )
your pay
( pool − operator's cut ) × your weight ÷ all members' weight

Every claim on this page, with the quote and the source behind it: the evidence.

Today

How collectives run today

Eight schemes pool publishers' content for AI companies, or pay publishers from a pool. Each is described here from its own documents.

Live · United States

Internet Licensing Collective (ILC)

ILC launched on November 12, 2026, with Yahoo Scout as its first licensee. An AI company takes one blanket license for a negotiated fee, which ILC doesn't publish, and reports its own use. Two terms favor the publisher: participation is non-exclusive, and a publisher can exclude a specific AI company. Uses are all or nothing. ILC's site tells publishers: "Choose which AI uses you authorize". Its participation agreement sets the uses in each AI company's license, and the check in ILC's publisher portal accepts only a license declaration, the file where a publisher states its terms, that grants every AI use, training included. The launch release says "There is no black box". The Royalty Distribution Policy that ILC shows publishers in its portal says the detailed calculations "are not publicly disclosed".

Live · the operator is the buyer

ProRata / Gist

ProRata's partner agreement puts "fifty percent (50%) of the Total Revenue" of its own answer products into a pool for all content partners, and ProRata's attribution engine splits it. No outside AI company pays in. Partner content "is never used for model training", ProRata says.

Launched in 2025 · the operator is the buyer

Perplexity Comet Plus

Perplexity announced Comet Plus in August 2025: subscription revenue goes to participating publishers, "minus a small portion for Perplexity's compute costs", split by "human visits, search citations, and agent actions". Perplexity runs the pool and is its only payer. Its launch partners were Condé Nast, Fortune, Le Figaro, Le Monde, the Los Angeles Times and The Washington Post. Perplexity's pricing page doesn't list Comet Plus. CNN's short agreement with Perplexity never became final, and CNN sued Perplexity in May 2026.

Offered · no licensee named · United States

CCC AI Systems Training License

The Copyright Clearance Center describes this license as "a voluntary, non-exclusive collective license" for training AI systems and for "the subsequent use of the trained system as a service". CCC tells rightsholders: "This license is open for your participation." Its public pages name no licensee, no price and no distribution rule for it.

Sign-up open · no license sold · United Kingdom

CLA, with PLS and ALCS

The UK's Copyright Licensing Agency is building a generative AI license with PLS, which represents publishers, and ALCS, which represents authors. CLA and PLS are regulated collectives. Publishers decide "the type of generative AI use they wish to license for", PLS says, and it names two: training and retrieval. Publishers that license through NLA Media Access can't join. No AI licensee has been announced, and the license terms sit behind PLS's member login.

Test with ProRata · Denmark

DPCMO

Denmark's collective for press publishers is regulated, and its members vote in a general assembly. Its one signed AI agreement is a framework with ProRata, which its transparency report calls a test; each member opts in on its own. In February 2026, DPCMO took OpenAI to court.

Live and paid · Norway · closed

Kopinor and the National Library of Norway

Kopinor licensed the National Library to train Norwegian language models on newspapers. The contract sets "an annual total remuneration of NOK 45,000,000 (in 2026)", the one published price among these schemes. Kopinor paid NOK 44.1 million to MBL, the Norwegian media businesses' association. It has one licensee and covers Norwegian press only.

Live · no revenue shown · Netherlands · closed

NDP Nieuwsmedia and GPT-NL

NDP Nieuwsmedia, the Dutch news publishers' association, licensed articles from more than 30 news titles, plus the news agency ANP, to TNO, a public research body, to train the GPT-NL model. Contributors who choose revenue sharing split 50% of the net revenue from licensing the model; the template also offers no compensation, or a one-time payment. "TNO will retain 50% of Net Revenues", says the published rule, and its half "will be re-invested" in the model. Each share is a token count times a quality multiplier. Publishers choose what to contribute, and TNO filters and weights it. Leaving takes six months' notice, and models already trained may be used for two more years. The published terms are TNO's template; NDP's own contract isn't public. No paying licensee of the model and no payout has been shown. The license covers NDP's Dutch member titles only.

Music has run on pools for decades: a streaming service is a licensing collective with a player on top. See how it maps onto FORA.

Elsewhere

Around the world

What publishers' bodies elsewhere are doing about AI licensing, one line each.

Europe

  • Live · music Sweden and Germany, music collectives. STIM runs a collective license for AI training and output; its first licensee is Songfox. GEMA licensed a training dataset to Klangio. It also publishes a licensing model at 30% of an AI provider's net income, with no licensee found.
  • Internal use only · test Finland, Kopiosto and a media pilot. Kopiosto's AI license covers a company's own staff, not AI developers. In a pilot, Finnish news media are "testing content use and pricing models with AI developers".
  • Announced Spain, CEDRO. CEDRO signed as the exclusive Spanish partner of the RSL Collective, a name ILC's partner page still uses. It has no license yet.
  • Complaint France, DVP and APIG. DVP and APIG each went to the competition authority in 2026 over Google's AI Overviews and AI Mode.
  • Complaint European Union, European Publishers Council. It filed an antitrust complaint over Google's AI Overviews. An Advocate General at the EU's top court says chatbot answers can infringe press publishers' rights.

Asia

  • Live Korea, Korea Press Foundation. It sells news data for AI training from the outlets it manages in trust, 97 of them in 2024. Each outlet sets its own price, and no buyer is named.
  • Announced Korea, Daum. The portal plans a pool of KRW 2 billion a year for partner outlets, split each quarter by how often its own AI summaries cite them.
  • Litigation Japan, the newspaper association. It says AI training and AI search need permission. Nikkei and Asahi are suing Perplexity, as is Yomiuri.
  • Litigation India, ANI v OpenAI. The Delhi High Court refused an interim ban; ANI has appealed. A government committee has proposed a mandatory training license paid into one collective.
  • Live Asia, LINC. LINC, a Singapore company, enrolls Asian publishers in ILC and other licensing channels. It names no AI licensee of its own.

Australia

The Americas

  • Litigation Canada, news publishers. They are suing OpenAI, where the Ontario court kept jurisdiction, and Anthropic.
  • Litigation Brazil, ANJ. ANJ and two other media associations won an injunction that bars Meta from using members' content for AI training. A Senate-passed AI bill would require pay for training.

Publishers' associations in Europe and Asia are asking AI platforms for transparency and equal treatment. The Zurich declaration of the German-speaking countries' publishers' associations says platforms "dürfen weder eigene Inhalte noch Inhalte einzelner Medien privilegieren": they may favor neither their own content nor that of individual outlets.

The scorecard

Eight questions the collectives answer differently

A question is in the table only if the eight schemes differ from each other on it. The last column describes a collective run on FORA.

QuestionILCProRata / GistPerplexity Comet PlusCCCCLA / PLSDPCMOKopinorNDP / GPT-NLOn FORA
Who pays you today? Yahoo ScoutProRata itself, from its own productsPerplexity itself, from subscriptionsNo licensee namedNo licensee announcedProRata, as a testThe National Library: NOK 45 million a yearTNO, from GPT-NL license revenue; no paying licensee shownNo one: no collective runs on FORA yet
Which uses do you license, and who picks them? Every AI use in each license; you can exclude a licensee or a sectionAnswers; no training; fixedPays on visits, citations and agent actions; choice not statedTraining; you opt into the licenseYou pick: training or retrievalRetrieval; you opt in per dealTraining only; fixedTraining only; fixedYou pick, per term; a pool leaves training out
Who picks whose content gets used, and is the method published? The AI company; method not publishedProRata's own engine; method not publishedPerplexity's own engine; method not publishedThe AI company, from content it already buys; method not statedYou choose what to offer; the AI company's pick is not statedProRata's engine, in the test; DPCMO's documents don't sayThe National Library "shall determine which publications are used"; method not publishedYou choose what to contribute; TNO filters and weights it for training, by a published ruleThe operator's search; the ranking rule is the operator's own, and it can publish it
Who counts the use your share rests on? The AI company, in its own reportProRata's attribution enginePerplexityNot stated for AI; licensee surveys in CCC's other licensesNot stated for AI; licensee reports, audits and surveys in CLA's other licensesProRata's attribution engineNo one: a lump sumTNO, once: your tokens in the training set, and their qualitySigned fetch records, with AI companies' reports checked against them
Is the split rule published? Principles; the calculations "are not publicly disclosed"The formula, not the attribution weightsThree categories of trafficNoGeneral splits; no AI rule yetNot onlineGeneral guidelines; the split among newspapers not foundYes, in TNO's template: token count times a quality multiplierEvery weight and threshold
What does the operator keep? Its costs, "projected at or below 10%"50%"a small portion"; 20% in press reportsNot published11% (CLA) and 6% (PLS) on standard licenses; not stated for AINot publishedNot published for this money50%, re-invested in the modelThe operator sets it, and publishes it
Is it regulated, and do members vote? Not regulated; no member vote; a publisher steering committeeA private company; no voteA private company; no vote foundNot regulated; no vote foundRegulated; publishers vote through trade associationsRegulated; members voteRegulated; member organizations voteNot regulated; no direct vote; a contributors' Content BoardNot regulated; each operator sets its governance
Does it reach copies an AI company took before any license? YesNo: retrieval onlyNot statedNot statedNo: "No retrospective cover"No; DPCMO is suing OpenAI over itThe library's own collections, yes; earlier training of its base models, noNo: it covers GPT-NL's own training onlyNo: a copy already held leaves no record to check
Who pays you today?
ILC
Yahoo Scout
ProRata / Gist
ProRata itself, from its own products
Perplexity Comet Plus
Perplexity itself, from subscriptions
CCC
No licensee named
CLA / PLS
No licensee announced
DPCMO
ProRata, as a test
Kopinor
The National Library: NOK 45 million a year
NDP / GPT-NL
TNO, from GPT-NL license revenue; no paying licensee shown
On FORA
No one: no collective runs on FORA yet
Which uses do you license, and who picks them?
ILC
Every AI use in each license; you can exclude a licensee or a section
ProRata / Gist
Answers; no training; fixed
Perplexity Comet Plus
Pays on visits, citations and agent actions; choice not stated
CCC
Training; you opt into the license
CLA / PLS
You pick: training or retrieval
DPCMO
Retrieval; you opt in per deal
Kopinor
Training only; fixed
NDP / GPT-NL
Training only; fixed
On FORA
You pick, per term; a pool leaves training out
Who picks whose content gets used, and is the method published?
ILC
The AI company; method not published
ProRata / Gist
ProRata's own engine; method not published
Perplexity Comet Plus
Perplexity's own engine; method not published
CCC
The AI company, from content it already buys; method not stated
CLA / PLS
You choose what to offer; the AI company's pick is not stated
DPCMO
ProRata's engine, in the test; DPCMO's documents don't say
Kopinor
The National Library "shall determine which publications are used"; method not published
NDP / GPT-NL
You choose what to contribute; TNO filters and weights it for training, by a published rule
On FORA
The operator's search; the ranking rule is the operator's own, and it can publish it
Who counts the use your share rests on?
ILC
The AI company, in its own report
ProRata / Gist
ProRata's attribution engine
Perplexity Comet Plus
Perplexity
CCC
Not stated for AI; licensee surveys in CCC's other licenses
CLA / PLS
Not stated for AI; licensee reports, audits and surveys in CLA's other licenses
DPCMO
ProRata's attribution engine
Kopinor
No one: a lump sum
NDP / GPT-NL
TNO, once: your tokens in the training set, and their quality
On FORA
Signed fetch records, with AI companies' reports checked against them
Is the split rule published?
ILC
Principles; the calculations "are not publicly disclosed"
ProRata / Gist
The formula, not the attribution weights
Perplexity Comet Plus
Three categories of traffic
CCC
No
CLA / PLS
General splits; no AI rule yet
DPCMO
Not online
Kopinor
General guidelines; the split among newspapers not found
NDP / GPT-NL
Yes, in TNO's template: token count times a quality multiplier
On FORA
Every weight and threshold
What does the operator keep?
ILC
Its costs, "projected at or below 10%"
ProRata / Gist
50%
Perplexity Comet Plus
"a small portion"; 20% in press reports
CCC
Not published
CLA / PLS
11% (CLA) and 6% (PLS) on standard licenses; not stated for AI
DPCMO
Not published
Kopinor
Not published for this money
NDP / GPT-NL
50%, re-invested in the model
On FORA
The operator sets it, and publishes it
Is it regulated, and do members vote?
ILC
Not regulated; no member vote; a publisher steering committee
ProRata / Gist
A private company; no vote
Perplexity Comet Plus
A private company; no vote found
CCC
Not regulated; no vote found
CLA / PLS
Regulated; publishers vote through trade associations
DPCMO
Regulated; members vote
Kopinor
Regulated; member organizations vote
NDP / GPT-NL
Not regulated; no direct vote; a contributors' Content Board
On FORA
Not regulated; each operator sets its governance
Does it reach copies an AI company took before any license?
ILC
Yes
ProRata / Gist
No: retrieval only
Perplexity Comet Plus
Not stated
CCC
Not stated
CLA / PLS
No: "No retrospective cover"
DPCMO
No; DPCMO is suing OpenAI over it
Kopinor
The library's own collections, yes; earlier training of its base models, no
NDP / GPT-NL
No: it covers GPT-NL's own training only
On FORA
No: a copy already held leaves no record to check

"Not stated" and "not found" mean the scheme's public pages don't say. These schemes compete as ways to pool content, and they can run side by side: six of the eight say participation is non-exclusive.

Beyond the table

What FORA adds

No scheme's public documents offer these.

  • A record of every fetch that the payer didn't produce. In the six schemes that say how use is counted, the AI company reports its own use, the operator or the licensee counts it, or no one counts per use. CCC and CLA haven't said for AI. On FORA, every fetch is a transaction for one URL. The operator's Exchange, the marketplace that lists members' content, signs the offer, and the AI company signs its acceptance.
  • Reports checked against the records, every period. The schemes' public terms offer at most an audit or a dispute right. On FORA, the operator reconciles each AI company's report with the fetch records before it splits the pool.
  • A share you can recompute. Seven of the eight don't publish enough to redo the sum. GPT-NL publishes its rule; its public documents don't say whether a contributor gets the token counts and quality scores the sum needs. On FORA, your statement gives your weight, the total weight, the pool, the operator's cut and the version of the rule.
  • Your own data, fetch by fetch. ProRata sends monthly statements down to the article; ILC sends periodic ones and makes them confidential. None promises data per fetch. On FORA, you see each of your transactions, and the report filed against it, as it arrives.
  • Your own priced terms beside the pool's. Join the pool with one term for everything, or a term per section and per use. Terms you price yourself run next to it, and the same fetch records back both.

One limit holds everywhere, FORA included. A record proves the fetch. What a model did with the page afterward comes from the AI company's report, and an audit checks it.

Selection

Who decides which publisher gets used

No scheme publishes how an AI system picks whose content to use in an answer. An AI company pays the same fee whichever publisher its system picks, so the split between publishers follows the pick.

Where the schemes say, the buyer decides. ProRata's and Perplexity's own engines choose the sources. ILC's public pages, its agreement and its policy don't say how a licensee chooses. GPT-NL comes closest, for training only: it publishes how TNO filters and weights each contributor's text. One contract names the picker:

The National Library shall determine which publications are used in the training of the various language models.

Kopinor and the National Library of Norway, training agreement

On FORA, the operator runs the search, and licensing is per URL. The operator writes the ranking rule, so the collective holds it and can publish it. Gutefrage, which runs FORA on helpster.de, uses the Elasticsearch search engine. The search only proposes URLs: each fetch is licensed on its publisher's terms, and the AI company's model still decides what to cite.

What FORA lacks

Where others are ahead of FORA

  • They run, and some pay. ILC has a licensee in Yahoo Scout. ProRata shares revenue with its partners under a public agreement. Kopinor has paid NOK 44.1 million to MBL. No collective runs on FORA yet.
  • A legal mandate, and enforcement. CLA, PLS, DPCMO and Kopinor are regulated: the law sets what they must disclose, and a supervisor oversees them. DPCMO is suing OpenAI, and ILC's participation agreement lets it sue unlicensed AI companies in a publisher's name. FORA is a protocol. It has no mandate, and it takes no one to court.
  • Copies AI companies already hold. ILC's license reaches content that AI companies already hold, "acquired before the RSL declaration was published", says its guide. RSL is the format ILC uses for that declaration. CCC, CLA, Kopinor and NDP license training. A pool on FORA licenses what leaves a fetch record, and a model trained on a copy it already holds leaves none.
  • A vote. DPCMO's members vote in a general assembly, as its transparency report sets out. FORA leaves governance to each operator.
Questions

Questions publishers ask

What does it cost to run a collective on FORA?

FORA is free. Each collective chooses how it prices its service.

Can I keep one AI company out?

Yes. Scope your terms to the AI companies you license. Any you leave out get no offer.

Does FORA hold the pool's money?

No. The operator collects the fees and pays publishers through its own payment provider. The protocol signs and records. It never moves money.

Can I stay in ILC and join a collective on FORA?

ILC allows it. Its FAQ says: "Participation is non-exclusive. You keep the right to license your content directly to anyone, and any existing licensing deals remain unchanged." If you want an ILC licensee to pay through a FORA collective, exclude it in ILC first.

Does a collective on FORA license training?

No. The pool licenses uses that start with a fetch, such as AI answers and search. A model that trains on a copy it holds fetches nothing, so no record exists to check its report against. You can still license training on your own priced term.

Why would an AI company join?

For one license to many publishers and one report format. It also gets licensed access to each member's pages, where an AI company outside the collective meets a per-fetch price or a block. No law makes it join.

Sources

Where each claim comes from

The quote behind each claim is on the evidence page.

All sources (31)

The Internet Licensing Collective

  1. ILC launches with Yahoo Scout as first commercial partner Internet Licensing Collective, press release
  2. For publishers Internet Licensing Collective
  3. Frequently asked questions Internet Licensing Collective
  4. Publisher Onboarding Guide Internet Licensing Collective
  5. Publisher Participation Agreement, version 1.0 Internet Licensing Collective No public URL: the agreement ILC's publisher portal asks publishers to accept
  6. ILC Royalty Distribution Policy Internet Licensing Collective No public URL: shown to publishers in ILC's publisher portal

ProRata and Perplexity

  1. Partner agreement for Gist ProRata.ai
  2. For publishers and creators ProRata.ai
  3. Introducing Comet Plus Perplexity
  4. Announcing Comet Plus launch partners Perplexity
  5. Pricing Perplexity
  6. How Perplexity's new revenue model works Digiday
  7. CNN sues Perplexity TheWrap

CCC, CLA and PLS

  1. AI Systems Training License: overview Copyright Clearance Center
  2. PLS offers new collective AI licensing opportunity to publishers Publishers' Licensing Services
  3. Distribution Charter Publishers' Licensing Services
  4. PLS opt-in licences: flyer Publishers' Licensing Services

DPCMO and Kopinor

  1. Transparency report 2025 (in Danish) DPCMO
  2. ProRata partners with DPCMO DPCMO
  3. A united Danish media industry takes OpenAI to court DPCMO
  4. Agreement on the use of copyright-protected material in the training of Norwegian language models Kopinor and the National Library of Norway
  5. Kopinor's largest payout ever (in Norwegian) Kopinor, via NTB

NDP Nieuwsmedia and GPT-NL

  1. Dutch news publishers contribute to developing GPT-NL IO+
  2. Content Contributor Agreement (template) TNO, GPT-NL
  3. Annex 2: Training Content Protocol TNO, GPT-NL
  4. Annex 3: Governance Charter TNO, GPT-NL
  5. Annex 6: Revenue Sharing Mechanism TNO, GPT-NL

FORA

  1. Licensing terms FORA documentation
  2. Transaction flow FORA documentation
  3. How money flows FORA documentation
  4. Protocol reference FORA documentation

Ask your collective these questions.

Take the scorecard's eight questions to any collective you're weighing. If you run a collective, see how one works on FORA.

The eight questions How money flows For Exchange operators