This page was translated automatically and has not been reviewed by a human. Some wording may be inaccurate.

Back to articles

Can media outlets escape AI and algorithms?

Artificial intelligence is flooding newsletters, filtering notifications, and summarizing WhatsApp messages. For media outlets seeking independence, it's back to square one: how do they maintain a direct connection with their audience when even alternative channels are coming under algorithmic control?

Aware that social networks are far from perfect partners, newsrooms have, in recent years, undertaken an ambitious project: building a direct and lasting relationship with their audience, despite an Internet dominated by a handful of companies. To circumvent the opaque filters of platforms, they have multiplied their efforts: investing in newsletters, developing their websites, testing WhatsApp or SMS… so many attempts to regain control over distribution.

Yet with the rise of artificial intelligence, these supposedly "alternative" channels are in turn coming under control. Automated summaries, reclassified notifications, reformulated content: widespread automation reshuffles the deck. Fleeting trend or lasting transformation, AI is once again upending editorial strategies. Newsrooms must, once more, reorganize themselves.

Newsletter: the end of a stronghold

For a long time, the newsletter was the preferred channel for editorial teams seeking independence. Freed from SEO constraints, they saw it as a direct, reliable tool capable of building a privileged relationship with their readers. For some media outlets, it even served as a springboard: like Vert in France, or Canadian media, which adopted it en masse after Meta blocked news on its platforms.

But this relative autonomy is now being challenged. According to Adweek, media outlets are already rethinking their newsletter strategy to adapt to the massive integration of AI into email services. Google and Apple are rolling out tools that automatically summarize, sort and prioritize emails. The result: the algorithmic logic that newsrooms were trying to avoid is now making its way into inboxes.

💡
To learn more about how Google's search engine is transforming with AI and its impact on publishers, read our article.

Notifications: the illusion of control

Some news organizations have bet on their own mobile application. Beyond providing a dedicated reading space, it allows them to send unfiltered

But here too, AI has muscled in. Smartphone makers are now offering automated notification summary systems, meant to simplify the lives of overwhelmed users. Apple was part of this… before abruptly backing down, and for good reason. The American company had been called out by the BBC, furious that Apple's AI had generated notifications falsely announcing Luigi Mangione's death or Rafael Nadal's coming out, while crediting the British media outlet and displaying its logo.

WhatsApp, another recently tapped channel, is no exception. Many media outlets have created groups there to transmit information directly to their community. But Meta, which is desperately trying to catch up with OpenAI, is determined to align all its services with AI, and its flagship application is no exception. The company recently announced that it would test conversation summaries powered by AI there.

Play with AI, without being overwhelmed by it?

Faced with this wave of automation, some media outlets are seeking not only to resist but to work with AI in hopes of maintaining some degree of control (more or less illusory). The Washington Post launched, for example, "Climate Answers", a conversational AI enabling users to explore its climate content. In France, Le Monde reached an agreement with Perplexity to integrate AI into its search engine, while Ouest-France developed "Topo", designed to answer readers' questions about the 24 Hours of Le Mans in order to "preserve its digital sovereignty".

💡
AI and media? Twice a month, in our 10 Links newsletter, we share our best reading on the topic.

Return to the web and physical: the last refuge?

For many media outlets, the real challenge remains becoming their own distribution platform again — rather than being mere content producers dependent on third parties to reach their audience.

This is what's driving a major comeback for website traffic, amid what Charlotte Klein calls an "apocalypse of referrals" in New York Magazine. Facebook, Google, Twitter (now X)… They're all now sending far fewer visitors to news sites.

Faced with this reality, some media outlets are rethinking their entire online presence. The Verge is a prime example: its editor-in-chief, Nilay Patel, describes the outlet—with a touch of irony—as "the last website on Earth." By introducing new features inspired by social networks, The Verge is positioning itself as a direct competitor to platforms like X and as a healthy destination in the midst of a polluted and frenetic web. The goal: to encourage readers to type theverge.com into the search bar rather than relying on an algorithm's recommendations.

On its site, The Verge organizes its articles into "Storystreams," feeds similar to those on social networks that aggregate articles, comments, social media posts, and more

This impulse toward taking back ownership also plays out offline. A trend exemplified, for instance, by WIRED and its recent initiative to let readers download and print their own micro-magazine on paper, or by Marsactu and its effort to provide access to its content via WiFi in cafés and libraries throughout Marseille.

While not completely abandoning the web, others are exploring more unexpected digital territories. Take Poetical, an editorial project that consists of inserting texts directly into users' digital calendars: essays, poems, quotes or article excerpts, transforming the personal calendar into a genuine reading space. The advantage of this strategy? It repurposes a purely utilitarian tool as a distribution medium, thus escaping the filters of traditional platforms to naturally slip into readers' daily lives.

Editorial teams could also play an educational role by explaining to their audience how to disable certain filters, use alternative tools, or configure their devices to receive information as it was intended.

But is it enough?

These alternatives, however inventive, remain marginal. Take calendars as an example: far less practical than email inboxes, they cannot handle the same volume of information. Where a user can subscribe to dozens of newsletters, it seems unthinkable that so many media outlets would clutter their personal calendar.

And above all: asking newsrooms to completely rethink their distribution model places the entire burden of a structural imbalance on their shoulders. How can they compete with the billions of dollars Silicon Valley invests in increasingly sophisticated systems and the rampant "crappification" of the internet? Especially since user behavior continues to concentrate on social networks, while website and app audiences slowly but steadily erode.

Regulate, rather than innovate alone

Beyond editorial or technical creativity, the stakes are fundamentally political and legal. Who owns the content used to train AIs? How can we ensure fair compensation for publishers? What transparency is needed for sorting and recommendation algorithms?

For now, major AI companies are advancing without significant obstacles, even winning their first legal victories, massively appropriating journalistic content for their training without compensation under the guise of "fair use" and taking advantage of a certain legal gray area that seems urgent to address.

AI is becoming a necessity rather than an option. "Opting out of AI is not easy," observes researcher James Jin Kang, who advocates notably for recognition of the right to refuse AI. "Governments, businesses, and communities must develop policies that not only regulate AI but also respect individual freedoms. Citizens must have the ability to live without AI without facing discrimination or exclusion from essential services". The same applies to access to information.

To go further