Meta has delayed its controversial new Meta business data terms following regulatory pressure in Britain and a court injunction in Norway. The rules, originally due to take effect on 30 October, would have allowed the company to crawl and scrape content from organisations with commercial accounts to feed automated systems.
Key points
- Meta postponed its commercial account updates after an antitrust inquiry in the UK and a court injunction in Norway.
- The proposed rules authorised automated bots to index, collect and analyse web pages linked to business profiles without payment.
- Publishers warned the harvested data could be used to train rival artificial intelligence services.
- Meta maintains the data collection was intended for advertising product recommendations rather than harvesting news.
Antitrust scrutiny in Britain
The decision to pause the rollout follows direct intervention by the UK Competition and Markets Authority. Officials opened an inquiry into whether requiring content companies to accept automated data collection constitutes an abuse of market dominance.
A formal investigation into Meta’s publisher terms launched by the British regulator is examining whether the contractual update creates unfair trading conditions. The authority said it was concerned that imposing the terms on media outlets reliant on Facebook and Instagram could break competition law.
Campaigners at the Movement for an Open Web submitted the complaint that triggered the British regulatory review. Tim Cowen, representing the group, said competition authorities must draw a firm line against unilateral data grabs by dominant technology platforms. He warned that without swift regulatory action, large tech firms would continue throwing their weight around against dependent publishers.
Legal injunction in Norway
Resistance to the contract changes mounted in Scandinavia earlier in the week. A Norwegian district court granted an emergency injunction on 28 October, temporarily barring Meta from imposing the conditions on 180 domestic media outlets.
The legal challenge was coordinated by the Norwegian Media Businesses Association, whose members represent roughly 97 percent of the country’s newspaper circulation. Lene Eltvik Vindfeld, chair of the association board, said: “We are satisfied that the court has agreed with us that Meta holds a dominant position in the relevant market and that the new terms are unreasonable.”
Randi S. Øgrey, chief executive of the association, said the financial foundation of independent reporting was at risk. She set out three demands from Norwegian publishers: explicit recognition of publisher opt-outs against web crawlers, separate voluntary agreements with remuneration before journalism is used in machine learning, and an assurance that non-consenting publishers will not be demoted across social feeds.
The dispute over artificial intelligence scraping
Under the contested text, commercial account holders had to permit automated software to index, cache, analyse, collect, retrieve or crawl web addresses associated with their accounts. The requirement applied automatically across all business clients, offering no mechanism to opt out while retaining access to commercial advertising and publishing tools.
The language caused alarm among legal specialists after a report by the Nieman Journalism Lab detailed how the clause could let Meta collect publisher data to train generative models. Legal experts consulted for that analysis concluded that the broad wording created a contractual licence for automated harvesting without compensation.
Meta subsequently removed the explicit word train from drafts of the agreement, but industry opposition continued. In the United States, Maria A. Pallante, president and chief executive of the Association of American Publishers, accused the company of abusing its dominant position to force consent from commercial partners.
Meta defends crawling practices
Meta insists that critics have misunderstood the practical function of its data collection. Andy Stone, a spokesman for the company, said on X that the updates were designed to extract accurate product listings from third-party retailers for on-platform advertising recommendations rather than indexing news organisations.
Stone said the company had paused the rollout to consult industry leaders. “We care deeply about our partners and do not plan to make policy changes without a full consultation with industry groups and leaders,” Stone said.
The delay comes as Meta faces broader financial scrutiny over its heavy capital expenditure on computing infrastructure, which recently saw Meta credit rating outlook cut to negative by credit analysts assessing its model training costs.
Standard consumer accounts remain subject to different policies. Meta already uses public posts, comments and media uploaded by regular personal accounts on Facebook and Instagram to train its foundational models, except where regional privacy frameworks restrict such processing.
Brussels and the regulatory outlook
The European Commission confirmed it had raised the commercial contract update directly with Meta. Sara Simonini, speaking for the European Union executive, welcomed the delay and noted that dominant gatekeepers under digital competition rules must provide balanced trading terms to business users.
The Commission is also checking how artificial intelligence companies collect training inputs as part of obligations under the EU AI Act. Regulators have requested technical summaries from model developers detailing the copyrighted material ingested by their systems. Meta has not yet set a new date for introducing revised commercial terms.
