Every morning, tens of millions of French-speaking internet users open a tab to find out what happened overnight. Web news, push notifications, personalized feeds: the sources are multiplying, but not all function in the same way. How do the main online information channels distinguish themselves in terms of format, personalization, and regulatory framework?
AI Aggregators and Online News Feeds: What’s Changing in Information Sorting
The homepages of major media outlets (20 Minutes, Le Monde, CNews) display a classic editorial selection, hierarchized by journalists. Google News offers an algorithmic ranking, filterable by theme and geographical area. In recent years, a third category has emerged: briefings generated or filtered by artificial intelligence.
Further reading : Discover the latest news and trends that are shaping our daily lives
Services like Microsoft Start or the summaries integrated into Perplexity AI produce daily news feeds personalized by AI, with automatic summaries and prioritization tailored to each user’s interests. Artifact, before announcing its closure in early 2025, offered a similar operation. This model differs radically from a traditional editorial feed, where the editorial team decides the order of topics.
The information published daily covers politics, current events, cultural life, as well as video games or music. To keep up with this thematic diversity, several internet users combine a generalist media outlet with a specialized aggregator. For example, there are sections dedicated to France, the world, health, or buying guides on the Communiqués du Net website, which gathers press releases sorted by category.
Read also : All the latest news and trends analyzed to better understand the world
Comparison of Online News Channels: Editorial, Algorithmic, and AI
The table below summarizes the structural differences between the three main types of channels used to consult the latest news online.
| Criterion | Editorial Media (20 Minutes, Le Monde) | Algorithmic Aggregator (Google News) | AI Briefing (Microsoft Start, Perplexity AI) |
|---|---|---|---|
| Topic Selection | Human editorial team | Algorithm based on popularity and location | AI model tailored to user profile |
| Personalization | Low (choice of sections) | Medium (followed themes, history) | High (custom summaries, adjusted tone) |
| Dominant Format | Long articles, video, live | Links to external sources | Short summaries, bullet points |
| Sorting Transparency | Displayed editorial line | Partially opaque criteria | Poorly documented recommendation logic |
| DSA Obligations (EU) | Classic editorial responsibility | Enhanced obligations (very large platform) | Enhanced obligations if threshold reached |
The “Sorting Transparency” column deserves special attention. Editorial media publish their editorial line, allowing readers to understand the angle of an article. Aggregators and AI tools, on the other hand, rely on recommendation systems whose parameters remain largely undocumented publicly.
Digital Services Act and Daily Consumption of Online News
The implementation of the Digital Services Act (DSA) of the European Union for very large platforms in 2023-2024, followed by its gradual extension to other players, changes the rules of the game. This regulation imposes transparency obligations on recommendation systems, content moderation, and combating misinformation.
In practical terms, a platform that aggregates news articles must now explain the main criteria of its ranking algorithm. It must also offer at least one sorting option not based on user profiling.
- Aggregators like Google News must document their recommendation parameters and provide an alternative chronological or thematic feed.
- AI briefing tools are subject to the same requirements once they reach the “very large platform” threshold defined by the DSA.
- Traditional editorial media remain subject to the editorial responsibility of press law, distinct from DSA obligations.
This regulatory framework has a direct impact on how the French access online news every day. The DSA creates a clear distinction between editorial responsibility and algorithmic responsibility, two regimes that now coexist on the same screen.
Agreements Between Media and AI Platforms: Tensions Over the Value of News
Since late 2023, several major press groups have engaged in negotiations, sometimes contentious, with developers of generative AI. The New York Times in the United States has launched a lawsuit. In Europe, publishers are negotiating licensing agreements to allow their articles to feed the responses of chatbots and AI aggregators.
This issue directly touches on the economic value of daily news. When an AI tool summarizes an article from Le Monde or 20 Minutes in three lines, the reader no longer necessarily clicks on the original source. Traffic to news sites decreases as AI summaries improve.
This transfer of value raises a fundamental question for anyone who consults web news daily. The apparent free nature of the AI summary relies on content produced by editorial teams that fund their staff through subscriptions or advertising. The licensing agreements currently being negotiated in France and Europe aim to restore a form of compensation, but their scope remains limited as long as the legal framework has not been stabilized by case law.
The online information landscape is thus evolving on two simultaneous fronts: European regulation, which governs algorithmic distribution, and the economic power dynamics between content producers and technology platforms. For the reader who opens their browser every morning, the choice of the news access channel determines as much what they read as what they will never see.