European publishers asked AI developers for transparency

FEP's 2023 position paper calls for training-data disclosure and licensing, not a ban on the technology.

Historical event

June 15, 2023

First source published

June 15, 2023

Site publication

September 18, 2026

The record

On 15 June 2023, at its General Assembly in Paris, the Federation of European Publishers, whose members are national publishers' associations across the European Union, adopted a position paper on artificial intelligence. The paper itself, dated 15 June 2023, addresses both the opportunities AI offers publishers and the copyright questions its use of published works raises. The Federation's own announcement confirms the paper was adopted by members "on the occasion of its General Assembly in Paris on 15 June."

What the documents show

The position paper states that AI "is not exempted from copyright rules, particularly when it uses pre-existing works," and it calls on AI developers to "guarantee the transparency of their training dataset, including regarding the works that were mined and where they were collected." It invokes the EU's 2019 Copyright in the Digital Single Market Directive, which lets rightsholders opt out of text-and-data-mining uses outside non-commercial research, and states that this framework is workable only if enforced with "stronger accountability by AI providers." The paper distinguishes an AI system used merely as an author's tool, which it says should not affect a resulting work's copyright eligibility, from output generated without a human's original creative choices, which it says should remain unprotected. It also flags a "data laundering" concern, where research conducted under a non-commercial exception is later folded into a commercial product without retraining on lawfully obtained data.

Why it stays interesting

The paper's own framing is instructive precisely because it is not a demand to stop AI development: it asks for licensing, disclosure and enforcement within an existing legal structure, the DSM Directive's opt-out mechanism, rather than a new prohibition. This is an editorial note: a reader comparing it with the individual author advocacy documents of the same period, such as membership-survey findings or model contract clauses, can see how an industry association's institutional position differs in tone and target, aimed at policymakers and AI developers rather than at publishers' own contract terms.

What to read next

The Society of Authors' 2024 member survey documents a different constituency's reported experience of the same period's AI developments, and reading it against this trade-association position paper shows how publishers' and authors' organisations were, at the same moment, making related but distinct arguments to different audiences.

• Which of the paper's requests concern the input side of AI, training-data use, and which concern the output side, generated content?
• How does the paper's account of the DSM Directive's opt-out mechanism compare with how that mechanism is actually described in the directive's own text?
• What would count, on the paper's own terms, as an AI developer meeting its call for training-dataset transparency?

A trade association's position paper records what its members were asking policymakers for at a stated moment, not a settled rule, and should be read as advocacy dated 15 June 2023 rather than as adopted law.

Sources & reading trail

The Federation's own position statement on AI, copyright, transparency and licensing.

Source published: 15 June 2023 · Retrieved: 16 September 2026

Confirms the paper was adopted by FEP members at its 15 June 2023 General Assembly in Paris.

Source published: Not established · Retrieved: 16 September 2026

Reports, filings, records and studies establish the entry; the reading of why it stays interesting is Chapter Ledger editorial analysis. This retrospective draft does not imply the site published on the event date.

Continue reading

• The EU's AI law requires a public summary of training content
• A membership survey counted authors' income lost to AI tools
• Publishers told the Copyright Office that licensing is the norm
• Browse the complete the ledger archive

Published September 18, 2026, not on the date of the event described.