Share

In-depth analysis of political news

©tungnguyen0905
News items published widely on social media are known to play a pivotal role in shaping political opinions. And yet, the algorithms driving these platforms often amplify content polarization. To enhance content diversity and foster a deeper understanding of the issues, CEA-List created an algorithm that enables automated news content analysis.

Developed as part of the BOOM (ANR) and AI4Media (Horizon 2020) projects, this innovative tool merges advanced language processing—including entity detection, theme analysis, and politician-centered sentiment classification—with political knowledge databases. It offers granular insights into media coverage, both traditional and social, including editorial biases, thematic framing, the evolving portrayal of public figures, and demographic biases.

When applied to French and Flemish (Belgium) media ecosystems, a pattern emerged: a persistent gender imbalance in political news, with women significantly underrepresented. Still, the average sentiment associated with female politicians was more positive than that of their male counterparts.

 

CEA-List analysis tool results: distribution of mentions and sentiment by gender in the most popular French media online ©CEA


 

Current research priorities now focus on four fronts: (1) quantifying selection bias to spotlight underrepresented perspectives on specific topics or events; (2) broadening sentiment analysis to include entities beyond politicians; (3) extending the framework to socially impactful areas; and (4) embedding the methodology into recommendation algorithms to diversify political news feeds—a critical step toward mitigating algorithmic polarization.

Key figure

62%

of French people get their news via social media*

  • *(Guénaëlle Gault, David Medioni, Survey “Les Français et la fatigue informationnelle”, L’ObSoCo, Fondation Jean-Jaurès, ARTE, june 2022)

Contributor to this article:

  • Evan DUFRAISSE, research engineer at CEA-List

Automated news characterization is a useful tool for navigating the complexity of opinions expressed in the press.

Rebecca Cabean

Evan Dufraisse

Research engineer — CEA-List

The tool provides journalists with an analysis that combines objective and subjective perspectives in news articles.

VRT (AI4MEDIA PROJECT PARTNER)

See also

Challenges

Artificial intelligence

From home to work, artificial intelligence has made in roads into virtually every aspect of our lives. It has transformed how we relate to others, do our jobs, and interact with the devices we use eve...
Read more
2024 Activity Report

Detecting texts generated by artificial intelligence

At the request of cybersecurity firm Vade, CEA-List engineered a first line of defense to new cyberattack vectors: AI-generated text detection.
Read more