
Video platforms and social networks have surpassed television as the primary source of news worldwide. According to the Reuters Institute’s Digital News Report 2025, 54% of respondents cite social media and video platforms as their source of information, compared to 52% for television and 51% for media websites and apps.
This shift redefines the pathways through which news reaches the public, and the implications for those who want to stay informed go beyond simply choosing a channel.
Algorithmic feed vs editorial line: what is changing in information sorting
The structural difference between a news feed on YouTube or TikTok and the front page of a news site lies in the selection mechanism. A media outlet applies an editorial line constructed by a newsroom. A recommendation algorithm optimizes the time spent on the platform, not journalistic relevance.
The Digital Services Act (DSA) has imposed a transparency obligation on very large platforms regarding the functioning of their recommendation systems since its implementation. In practice, the descriptions provided often remain too vague for a user to understand why certain content is suggested to them over others.
For those who wish to follow the news daily without relying on a single channel, it remains useful to cross-reference approaches: aggregators, thematic newsletters, and reference media. In this regard, visiting the Recommandons site allows access to a structured selection of sources covering politics, culture, science, and practical life.

Trust in the media in France: the figures from the Reuters Institute
The French context presents a unique profile. Trust in information peaks at 29%, below the global average. Only 36% of French people claim to be interested in the news. These two indicators, from the same report, depict an audience that does not reject information but doubts its reliability.
This distrust does not only affect traditional media. It extends to content shared on social networks, where the virality of a post guarantees no editorial verification.
What distrust concretely produces
A distrustful reader adopts two opposing behaviors. Either they gradually turn away from the news (the phenomenon of “news avoidance” documented in the Reuters report), or they multiply sources without hierarchy, leading to information overload. Both result in the same outcome: a fragmented understanding of substantive issues.
The answer does not lie in a nostalgic return to the television news. It involves a discipline of source selection, identifying those that practice verification and methodological transparency.
Social networks and information for young people: high usage, limited trust
Young audiences perfectly illustrate this paradox. They massively consume news via Instagram, TikTok, and YouTube, but several recent studies show that they are wary of it. A survey reported by the Swiss press confirms that young people get their news from social media but do not trust it.
This gap between usage and trust creates space for hybrid formats: news podcasts, editorial newsletters, YouTube channels for analysis led by identified journalists. These formats retain the accessibility of platforms while reintroducing an editorial signature.
- YouTube and TikTok serve as gateways to news, but content is rarely contextualized beyond a few dozen seconds.
- Thematic newsletters (politics, science, culture, finance) allow for receiving a selection filtered by an editor, with a weekly or daily rhythm suited to the subject.
- Aggregators like Google News rank articles based on relevance, reliability, and timeliness, but their ranking criteria remain partially opaque despite the DSA.

AI Act and algorithmic transparency: the European regulatory framework redefining access to information
The European regulation on artificial intelligence (AI Act) adds an additional layer to transparency obligations. Since August 2026, companies using generative AI systems must clearly indicate content produced by artificial intelligence. This obligation covers automatically generated texts, images, and videos.
For a reader, this regulation has a direct consequence: synthetic content must now be identifiable. In theory, an article or video produced by AI and distributed on a European platform will carry an explicit mention. In practice, enforcement remains uneven across platforms and countries.
DSA and recommendation systems
Article 38 of the DSA goes further by requiring very large platforms to offer at least one recommendation option not based on profiling. In concrete terms, a user should be able to access a chronological news feed, without algorithmic personalization. This option already exists on some platforms, but it is often buried in the settings.
We recommend systematically activating these options when available. A chronological feed exposes users to a diversity of topics that the personalization algorithm tends to reduce in favor of already consulted themes.
Building an effective news monitoring system in 2026
Staying informed without being overwhelmed requires a balance between depth and coverage. Three principles structure a functional monitoring system:
- Limit sources to a maximum of ten, covering politics, economy, science, and culture, prioritizing those that publish corrections and updates.
- Distinguish discovery channels (social networks, aggregators) from in-depth reading channels (news sites, long newsletters, podcasts).
- Set fixed times to check the news rather than being subjected to continuous notifications, which reduces information fatigue without creating gaps in coverage.
The shift to platforms as the primary source of information is irreversible. The European regulatory framework attempts to mitigate its negative effects, but the responsibility for filtering, verifying, and prioritizing largely falls on the reader. Choosing sources with the same rigor as one chooses work tools remains the most reliable method to follow the news without being overwhelmed.