Meta Turns Its AI Assistant Into a News Gateway With New European Publisher Deals

When users in Europe open WhatsApp or Instagram and ask Meta’s AI assistant what is happening in Madrid, Paris or Berlin, the answers now increasingly come with a byline.

Under a series of deals announced in March, Meta AI can pull from stories by News Corp, Spain’s Grupo Prisa, France’s Le Figaro and Germany’s Süddeutsche Zeitung, summarize the reporting and serve it directly inside Meta’s apps—complete with the name of the outlet and a link to the original article.

The expansion marks a new phase in Meta’s relationship with news: after years of downplaying journalism in its social feeds, the company is paying major publishers to supply the fuel for its most important new product.

Meta detailed the latest move in a March 13 post on its corporate site titled “Bringing More International News and Content to Meta AI.” The company said it is “beginning to offer a wider variety of real-time content on Meta AI—from global, breaking news to entertainment, lifestyle stories, and more.”

“When you ask Meta AI questions about the news, you’ll receive information and links that draw from more diverse content sources,” the post said. Meta said the integrations are designed to “facilitate easier access to information by linking out to articles, allowing you to visit these partners’ websites for more details while providing value to partners, enabling them to reach new audiences.”

The announcement, published on Meta’s Europe, Middle East and Africa newsroom, named News Corp, Prisa, Le Figaro and Süddeutsche Zeitung as new or expanded content partners for Meta AI in Europe.

From Reuters to News Corp: Building an AI news backbone

The European deals sit on top of a global licensing strategy Meta has been assembling over the past year and a half.

In late 2024, Meta signed a multiyear agreement with Reuters allowing its AI systems to use the news agency’s reporting to answer user questions about current events, with short summaries and links to full articles. That arrangement was an early model for how Meta AI would handle labeled, real-time news answers.

On Dec. 5, 2025, Meta announced commercial AI data agreements with a slate of major outlets, including USA Today and the USA Today Network, CNN, Fox News, Fox Sports, People magazine’s parent company, Le Monde Group in France, and conservative titles the Daily Caller and Washington Examiner. Under those deals, Meta said, its assistant could draw on partners’ content to provide “real-time answers” about news and current events and would show the source and a link when it did so.

Then, in early March this year, Meta and News Corp unveiled one of the largest of these arrangements to date. The multiyear deal is worth up to $50 million a year for at least three years, according to people familiar with the terms, implying a total value of as much as $150 million.

The agreement covers News Corp’s U.S. and U.K. news brands, including The Wall Street Journal, Barron’s, the New York Post, The Times and The Sun, along with their archives. It grants Meta rights both to train its AI models on News Corp content and to retrieve that material in real time to answer user questions inside Meta AI.

Australian titles such as The Australian are not part of this specific pact, News Corp has said.

At an investor conference discussing the Meta deal, News Corp chief executive Robert Thomson described his company’s role in stark terms. “We’re essentially an input company,” he said, likening the publisher’s journalism to semiconductors, data centers and energy as essential inputs to the AI economy. Thomson has framed News Corp’s broader strategy as “woo or sue”: reach paid agreements with AI firms or pursue litigation when they use news content without authorization.

Europe’s flagship dailies plug into Meta AI

The March 13 announcement extends this model across some of Europe’s most influential print titles.

In Spain, Grupo Prisa—owner of El País and other news and radio brands—said Meta had “announced an alliance” with the group, along with News Corp, Le Figaro and Süddeutsche Zeitung, to broaden the reach of their news and content through Meta AI in Europe. Prisa and Meta both emphasized that answers provided by the assistant will include links back to partner websites.

Meta’s Spanish-language newsroom entry, under the headline “Meta se asocia con PRISA, News Corp, Le Figaro y Süddeutsche Zeitung para ampliar las noticias y el contenido en Meta AI en Europa,” stressed that the assistant would offer more real-time news, entertainment and lifestyle content to European users.

French- and German-language posts highlighted the participation of Le Figaro and Süddeutsche Zeitung and framed the move as an expansion of Meta AI’s news offering in Europe. None of the regional announcements disclosed financial terms or detailed exactly which content types—such as video, photos or deep historical archives—are included.

Based on patterns in other AI licensing agreements, industry lawyers say such deals typically bundle rights for training models on past content and for using ongoing feeds to generate summaries and answers, as long as the AI cites the original outlet and links back to it. Neither Meta nor the European publishers have publicly confirmed those specifics.

How Meta AI now delivers the news

Meta AI is embedded into the search bars and messaging interfaces of Facebook, Instagram, WhatsApp and Messenger, and is also available through a standalone web interface and on devices such as Ray-Ban Meta smart glasses.

When a user asks a question that requires up-to-date information—for example, “What did the French government announce about pensions today?”—Meta AI draws on its licensed feeds and other sources to generate an answer. For stories that rely on partner content, the assistant identifies the publisher in the response and includes a link to the relevant article on that outlet’s site.

Meta has said its goal is to make the assistant more responsive, accurate and balanced on fast-moving events by incorporating a wider range of professional news sources. It also presents the linking as a benefit to publishers that can “reach new audiences” inside Meta’s apps.

The experience is different from Meta’s earlier news products. Rather than scrolling a feed of links or visiting a dedicated News tab, users ask for information directly and receive a synthesized explanation—with the option, but not the requirement, to click through to the full story.

A new business model for publishers

For publishers, the deals offer a mix of opportunity and risk.

Large media groups have been under sustained financial pressure as digital advertising has concentrated around a handful of tech platforms and subscription growth has slowed. AI licensing presents a new revenue stream that, at least in the short term, can be more predictable than pageview-driven ad sales.

In a joint statement with Meta last year, USA Today’s parent company called its multiyear AI licensing partnership “a responsible collaboration between trusted journalism and the information ecosystem.” The company said answers in Meta AI would include “information and links” from USA Today’s reporting.

Trade groups representing news organizations have argued that such arrangements are essential if AI companies are going to use journalism in their products. The News/Media Alliance, a U.S.-based association of publishers, has said “responsible AI starts with licensing” and warned that unlicensed scraping of articles for AI training poses a threat to the livelihoods of journalists and media workers.

At the same time, some media executives and researchers worry that AI assistants could erode the very traffic that underpins many outlets’ businesses. If users feel they already have the gist of a story from a chatbot summary, they may be less likely to click a link, subscribe or see the advertising that funds reporting.

Academic work has suggested that when publishers block AI crawlers, they can lose a significant share of web traffic, hinting at the leverage large AI and platform companies may ultimately hold in these negotiations.

Critics also point to a growing divide between major national and international brands, which can command eight- or nine-figure AI licensing agreements, and smaller or local outlets, which often struggle to get similar deals. There is concern that as AI assistants increasingly route attention and revenue toward a curated set of “authoritative” partners, the rest of the media ecosystem may be pushed further to the margins.

From news retreat to AI resurgence

The pivot to AI-powered news comes after Meta spent much of the past decade dialing down the role of journalism on its core platforms.

Facebook launched a third-party fact-checking program in 2016 under pressure to curb misinformation. But starting in 2021, the company began de-emphasizing political and news content in users’ feeds. By 2024, it had shut down its dedicated Facebook News tab in the United States, Australia and several European countries and wound down many direct-payment programs for publishers.

In Canada, Meta went further. In August 2023 it blocked users from viewing or sharing any news links on Facebook and Instagram in response to the country’s Online News Act, which aimed to force large digital platforms to compensate news businesses. The restrictions, which remain in place, drew criticism from the Canadian government and press freedom advocates, particularly during emergencies such as wildfires when access to information is critical.

At the same time, Meta has invested heavily in artificial intelligence, developing its Llama family of models and reorganizing its AI teams under a “superintelligence” unit. Meta AI is now integrated across its products and available in more than 200 countries, positioning the assistant as a central intermediary for searching and discovering information.

Legal and regulatory pressure has been mounting around how AI firms obtain the data they use. Meta and other technology companies face lawsuits from authors, media organizations and other rights holders who say their works were copied without permission to train AI models. In the United States, a federal judge last year dismissed some claims against Meta, but litigation continues in other jurisdictions, including France, where publishers have accused the company of economic “parasitism” over AI training.

Against that backdrop, the move to high-profile licensing deals with organizations such as Reuters, News Corp, Prisa and Le Monde allows Meta to argue that it is willing to pay for content while it continues to fight broader legal challenges.

A new front door to information

For users, the practical effect of Meta’s latest agreements is that more of the world’s news will be encountered not as a homepage or a social link but as a paragraph in a chatbot window, stamped with a source and an invitation to read more.

For publishers, the arrangements promise cash and exposure but create deeper dependence on a platform that has in the past changed course abruptly on news distribution and payment.

For societies and regulators, the deals raise questions about who will shape public understanding of events as AI assistants become a default way to ask what is happening in the world—and about how transparent companies such as Meta will be about whose voices are amplified when the answers appear.