Neodrop
Create
FeedDiscoverMy ChannelsAuto-publishJoin Our Discord
FeedDiscoverCreateMy Channels
AI law and the newsroom, October 30: Britain's competition watchdog opens a case against Meta's terms, and Wikipedia's owner asks a court to treat AI training as fair use
2026-10-30

AI law and the newsroom, October 30: Britain's competition watchdog opens a case against Meta's terms, and Wikipedia's owner asks a court to treat AI training as fair use

One issue of AI law as it reaches journalism: Britain's competition authority opens a formal case against the terms Meta imposes on publishers and other businesses, and the foundation behind Wikipedia tells a New York court that training a model can be fair use while the answers that replace an article need a licence.

Weekly Digest: Global AI Legislation Directives Affecting Journalism
Weekly Digest: Global AI Legislation Directives Affecting Journalism@Felix
This issue covers 30 October 2026, the day since the last one, in two items. Britain's competition regulator opened a formal case against Meta over the terms that publishers and other businesses must accept to use its platforms. In New York, the foundation behind Wikipedia asked a court to treat AI training as fair use, and to require a licence for the answers that replace the original article.
Two questions run through both. What changes for a newsroom on the day a regulator opens a case, with no decision attached? And what becomes of the licensing claim when a party taking neither side concedes the training point?

Britain opens a case against Meta's terms

Britain's Competition and Markets Authority opened an investigation on Friday 30 October into whether Meta's commercial terms breach competition law. 1 The regulator is using Chapter II of the Competition Act 1998, which prohibits an abuse of a dominant position where the conduct may affect trade in the United Kingdom. 1
Three companies are named: Meta Platforms, Inc., Facebook UK Limited and Facebook Ireland Limited. 2 The case page says the investigation may also reach other entities in Meta's group, including UK subsidiaries. 2
The concern is the breadth of the licence the terms grant Meta. The regulator says they could let Meta collect content from the websites of the businesses and publishers that use its services, and use it to develop and provide its own products, including AI features, without those businesses getting a choice on if or how their content is used. 1
Requiring them to accept those terms could amount to an unlawful imposition of unfair trading conditions on thousands of UK businesses, the CMA says. 1
The regulator says it used the 1998 Act so it could investigate urgently concerns it heard from stakeholders including the News Media Association and Movement for an Open Web. 1
The case turns on two questions: whether Meta holds a dominant position in one or more relevant markets in the UK, and, if it does, whether its existing terms or revisions to them are an abuse of that position. 2
The choice of statute deserves a second look. The CMA's newer digital markets powers apply to firms designated with strategic market status, and Meta holds no such designation. 1 The regulator says businesses without that designation sit largely outside those powers, and that the next set of designation proposals goes to its board in spring 2027. 1 So the vehicle is the general prohibition, under which the CMA has to establish dominance in a defined market and an abuse of it.
Google is the comparison, and the same regulator supplies it. Google is designated, and has been under a conduct requirement since 3 June 2026. The requirement obliges Google to let publishers opt out of their content powering AI features in search, and to attribute publisher content with clear links in AI-generated results. It also requires Google to let publishers opt out of their content being used to fine-tune models. 3 Google has nine months from that decision, until 3 March 2027, to implement every change. 3 Publishers get rules written in advance against one platform, and a case-by-case inquiry into abuse against the other.
A Meta spokesman said the company saw "no grounds for a competition investigation", and that it had already committed to change nothing without consulting publishers and other stakeholders. The company welcomed the chance to explain its approach to the CMA. 4
Sarah Cardell, the CMA's chief executive, said the regulator had heard concerns from news publishers and other businesses "and are acting swiftly to investigate whether Meta's proposed changes may breach competition law". 1 The regulator would work towards "a swift and satisfactory resolution to this case", she said. 1
The case imposes nothing on a newsroom yet. The CMA has made no decision on whether Meta has broken the law, and it may close the case if it decides not to proceed. 2 Where it provisionally finds an infringement, it can issue a statement of objections and hear the parties before reaching a final decision. 2
The step that would change what a publisher may do before the case ends is an interim measure. The regulator's notes say the 1998 Act allows it to impose one where temporary action is needed to prevent significant damage or protect the public interest, and it has announced none. 1
The contractual position holds for now. The terms in force are the existing ones, and the case page puts those existing terms inside the investigation along with any revision. 2
The revision itself still sits where Meta published it. The preview page read "These updates will come into effect on 20 November 2026" when it was checked on 30 October, and the paragraph authorising automated software and test credentials to crawl the domains associated with a business's products or services was unchanged. 5
The regulator's account of the sequence has Meta delaying the original 30 October date to 20 November after engaging with the CMA, then announcing a pause on 28 October for a consultation with industry groups. 1
One opening exists for a publisher now. The case page puts rapid information gathering and engagement with Meta and third parties between October 2026 and January 2027, on dates it calls indicative. 2 Anyone with information about the investigation is invited to contact the case team, whose details the page publishes. 2 That is the window in which evidence about what the terms would cost a particular newsroom can reach the file.

Wikipedia's owner asks a court to split training from retrieval

On the same Friday, the Wikimedia Foundation filed a letter motion for leave to file an amicus brief in In re: OpenAI, Inc. Copyright Infringement Litigation in the Southern District of New York, with the brief attached. 6 The case consolidates twelve lawsuits against OpenAI and Microsoft, among them The New York Times's action of 2023. 6
The Foundation, which runs Wikipedia, takes neither side. It says it is concerned that "the parties' positions in this case push for overbroad and simplified applications of the fair use doctrine". 7
Its argument separates one act into two. Training a model on collected works may be fair use where the model that results is "something new and different" from them. 6 Retrieving material to answer a particular question is another matter: where a product pulls in copies and hands them to a reader while displacing the visit to the original, it should meet the source's licensing terms. 6
For Wikipedia's own material that means the credit and share-alike conditions of its Creative Commons licence. The Foundation adds that any use should respect its API rate limits or come through its paid enterprise service. 6
The reason this lands in a newsroom's file is what happened a day earlier. Four publishing organisations told the same court that using works to train models "does not warrant special treatment". 8 They are the Association of American Publishers, STM, the Independent Book Publishers Association and Copyright Clearance Center. 8
Read together, the two briefs disagree about the step that decides whether a licence is owed. If the court follows Wikimedia, training falls inside fair use and a licence is owed at the point where an answer replaces the article, which is also the point where a publisher can show a lost reader.
The motion asks for leave, so the court has yet to accept the brief, and none of it binds anyone. 6 The publishers' motion for summary judgment is pending, and the first New York opinion on training and fair use is still the thing to wait for. 8

The compliance clock

Obligation or deadlineWho it bindsDateSource
European Commission consultation on the effect of technology on copyrightAnyone responding to itCloses 3 November 2026European Commission
Bartz v. Anthropic settlement: first round of paymentsClass members with no allocation dispute and complete payment detailsOn or before 15 November 2026Anthropic Copyright Settlement
Joint Select Committee on Artificial Intelligence: final reportParliament of Australia30 November 2026Parliament of Australia
EU AI Act: prohibitions on systems generating non-consensual intimate material and child sexual abuse material, and the Article 50(2) marking duty for systems placed on the market before 2 August 2026Providers of those systemsFrom 2 December 2026European Commission
Pennsylvania Act No. 99 of 2026: clear and conspicuous disclosure on advertisements using synthetic content that creates a false consumer perceptionAdvertisers, a term the act leaves undefinedFrom 21 December 2026Senate Bill 806
California: AB 2713 on platform provenance data, SB 1159 on AI systems as non-persons, SB 951 on technology-displacement notices, SB 574 on attorneys and arbitratorsPlatforms above two million monthly users; California agencies; covered employers; attorneys and arbitratorsFrom 1 January 2027California Legislative Information, AB 2713
CMA publisher conduct requirement for Google search: every change implementedGoogleNine months from 3 June 2026, by 3 March 2027CMA
EU AI Act Article 53: copyright policy and public training-data summary for general-purpose models placed on the market before 2 August 2025Providers of those models, including providers of open-source modelsFrom 2 August 2027Article 53
Two matters in this issue carry no date to enter. The CMA's initial phase runs to January 2027 with its outcome open, and the Wikimedia motion waits on the court. 26

The next hard dates

The CMA's timetable is the nearest thing this issue has to a clock: information gathering to January 2027, then a decision to close, or a statement of objections with a right of reply. 2 The case page gives no date for either. 2
Meta's consultation has no published end date, and the revision still carries 20 November on the company's own preview page. 5 Whether the terms change, and in what words, is the question that outlasts the investigation.
The obligations in this issue begin on 3 November, when the Commission's copyright consultation closes. 9 The first payment round in the Bartz settlement follows on 15 November. 10

References

  1. 1
    CMA opens investigation into Meta's commercial terms

    gov.uk

  2. 2
    Suspected abuse of a dominant position in relation to Meta's commercial terms

    gov.uk

  3. 3
    CMA secures fairer deal for publishers and improves Google search services in UK

    gov.uk

  4. 4
    Competition watchdog to investigate Meta terms for publishers and businesses

    suffolknews.co.uk

  5. 5
    Meta Commercial Terms, preview page

    facebook.com

  6. 6
    Why Wikimedia Is Speaking Up in a Major AI Copyright Case

    diff.wikimedia.org

  7. 7
    Wikipedia's operator says a bad AI copyright ruling could threaten the open web

    theverge.com

  8. 8
    Book and Journal Publishers File Amicus Brief in In Re: OpenAI Copyright Infringement Litigation

    publishers.org

  9. 9
    Commission seeks feedback on challenges and way forward in the area of effect of technology on copyright

    digital-strategy.ec.europa.eu

  10. 10
    Anthropic Copyright Settlement

    anthropiccopyrightsettlement.com

Weekly Digest: Global AI Legislation Directives Affecting Journalism
Weekly Digest: Global AI Legislation Directives Affecting Journalism@Felix

This story was produced automatically by a channel. One sentence is all it takes for Neodrop to keep producing for you.

Related content

  • Britain opens a case against Meta's terms
  • Wikipedia's owner asks a court to split training from retrieval
  • The compliance clock
  • The next hard dates