Opening an app on your phone and coming across a news video without having searched for it: this situation has become the default mode of information for a large part of the public. The shift from a voluntary search for information to a discovery driven by algorithms reshapes how everyone constructs their worldview. Understanding these mechanisms allows us to take control over what we read, listen to, and watch on a daily basis.
Algorithmic Discovery and Online Information: A Change in Logic
For a long time, staying informed required an active gesture. You would buy a newspaper, turn on the radio at a specific time, or type a keyword into a search engine. This pattern is receding.
According to the Digital News Report 2025 from the Reuters Institute, weekly consumption of news videos on social media reached 65% in 2025, compared to 52% in 2020. Information now arrives in a continuous stream, mixed with entertainment, product recommendations, or creator content.
This model of algorithmic discovery means that the platform partly chooses the topics for you. The main risk: seeing your informational universe shrink without realizing it. An algorithm optimizes engagement, not the diversity of viewpoints. To counter this effect, one must combine this passive exposure with active approaches, which we detail further on.
Portals like goinformation.info gather diverse news feeds to allow quick access to multiple sources from a single entry point.
News Creators and Traditional Media: Where to Place Your Trust
You may have noticed that a creator on TikTok or YouTube can summarize a geopolitical event in three minutes, while a TV news segment dedicates thirty seconds between two reports? This format particularly appeals to those aged 18-24.
Trust is shifting from the media to the individual. This shift poses a concrete problem. A creator does not have a newsroom behind them, no displayed ethical charter, no proofreader. Their credibility relies on their personal reputation. In contrast, a structured media outlet applies (in principle) verification rules before publication.
Combining the Two Without Drowning
The solution is not to choose a side. It consists of cross-referencing formats:
- Follow two or three creators specialized in topics that interest you (geopolitics, economics, sciences), ensuring they cite their sources in the description or comments.
- Keep at least one generalist reference media for topics outside your usual radar, whether it’s an online newspaper, a news channel, or a radio show.
- Occasionally consult a feed aggregator (RSS or dedicated app) to spot angles that neither your creators nor your main media cover.
A creator who never cites their sources is a warning sign, regardless of their number of followers.
Generative AI and News: An Entry Point to Verify
Conversational assistants (ChatGPT, Gemini, Copilot, and others) are beginning to serve as access points to news. The Generative AI and News Report 2025 from the Reuters Institute indicates that the use of generative AI for information has risen from 3% to 6% in the studied markets. The figure remains modest, but the trend is clear.
Why is this caution justified? An AI assistant does not produce reporting. It reformulates and synthesizes existing content. Three reflexes allow you to use this tool without getting trapped:
- Check if the response cites identifiable primary sources (media name, publication date). A response without a source is no better than a rumor.
- Compare the update date of the response with the event in question. Some models work with data that is several months old.
- Directly consult the original article to which the AI refers. Reformulation can distort a nuance, reverse a conclusion, or omit context.
Generative AI acts as an additional filter, not as a source. Always trace back to the original article remains the only guarantee of reliability.
European Regulation and Pluralism of Online Media
The European Media Freedom Act, which has come into effect in the European Union, aims to protect the editorial independence of media and regulate the concentration of press groups. This regulatory framework requires platforms to be more transparent about how they prioritize news content.
For the reader, this changes little in the immediate daily context. The effect will be measured in the medium term: platforms will have to explain why one piece of content appears in your feed rather than another. This obligation for algorithmic transparency could eventually allow for better customization of information preferences.
What This Means in Practice
European media benefit from a legal foundation to resist commercial or political pressures. For a French-speaking reader, this means that content produced by newsrooms subject to this framework offers an additional guarantee of independence compared to sources hosted outside the EU.
Favoring identified media, cross-referencing at least two sources on a sensitive topic, and tracing back to original articles when going through an intermediary (social network, aggregator, AI): these three habits are enough to build a solid information watch. The time spent verifying information is always better invested than the time spent consuming ten pieces without perspective.
