UK watchdog locks horns with Meta over fears Zuck could scrape British websites to train AI
Peers urge government to “use all the powers at their disposal” to stop Meta using publishers’ content for AI development.
The UK government has confirmed that the British competition watchdog is engaging directly with Meta over controversial new terms that could allow the tech giant to crawl British businesses' websites for AI development.
Speaking during a House of Lords debate yesterday, government minister Baroness Ramsey of Wall Heath said that the Competition and Markets Authority (CMA) was engaging directly with Meta "to establish the facts."
"The CMA recognises the urgency of this matter, particularly for businesses that rely on Meta services to reach their customers," she said.
"A delay does not resolve the questions raised by the effect of the new terms."
An initial version of Meta’s updated terms allowed it to use “any content” from websites linked to business pages to “provide, develop, improve, research, train, and maintain the integrity” of its products and services.
It has now revised the terms to remove the word “train” and no longer refers to “any content.” However, the rules still allow Meta to “periodically pull updates” from business users’ websites to develop, improve and research its advertising and business products.
Meta says the practice is intended to help businesses create and deliver advertising, promote their products and services, and reach potential customers.
Meta has delayed implementing its updated commercial terms from October 30 to November 20 after publishers and other businesses objected.
Liberal Democrat peer Lord Clement-Jones claimed that Meta had pushed back the rollout of its new policies without fully changing their substance.
"It does not use the word 'train' any more, but it can still crawl publishers’ websites to 'develop and improve' its products," he said.
He also asked whether the terms let Meta bypass paywalls and technical protections.
Scraping by
Lord Brennan of Canton urged the government to intervene against what he described as "a threat to trusted journalism."
"I urge the Minister to be robust on this matter and the Government to use all the powers at their disposal," he added.
Baroness Ramsey also said that removing explicit references to training had failed to resolve the dispute.
"The removal of the word 'train' has not resolved the concerns raised about the revised terms, particularly on whether the remaining provisions could still permit content to be used for AI training," she said.
READ MORE: "Don't record me!": Meta wants to stop Glassholes prying, perving and pirating
Meta has disputed that interpretation and maintained that website owners can use robots.txt to communicate their preferences about automated crawling.
The minister also confirmed that the European Commission received a formal competition complaint about Meta's revised terms on October 22.
No formal enforcement action has been announced.
Machine understands that Meta's revised terms are not intended to enable general AI training or the scraping of news websites, but to improve its advertising products.
The aim is to obtain accurate information about products for sale on third-party websites to improve advertising and content recommendations on Meta's platforms.
A Meta spokesperson said: "We've heard publishers' feedback on our proposed ads and business terms update.
"We care deeply about our partners and do not plan to make policy changes without a full consultation with industry groups and leaders.
"We're committed to getting this right, and as we expect this consultation could extend beyond November 20, in which case we’d delay implementation of the updated terms. Until then our existing terms remain in place."
The CMA is expected to provide an expedited written response to the House of Lords Communications and Digital Committee setting out its position.
Meanwhile, the UK government is reviewing mechanisms to give creators greater control over how AI development uses their work, with findings expected in the first part of 2027.
