Publication Date: March 2024
Journo News Monitor: an AI-powered study on Turkey’s news media

An AI-powered study on Turkey’s news media, credits Journo News Monitor

The Journo News Monitor collected data on Turkish news throughout 2023 using two dashboards: one examining news-related Google search results and the other focusing on the content of news websites in Turkey. After a year-long study, Journo’s editor Emre Kizilkaya analyzed the findings in this report. The analysis, utilizing AI, highlighted a bias in Google’s algorithm favoring pro-government media, evident from these outlets being prioritized in 74.8% of news-related searches. There was also a 14% decline in the diversity of news sources recommended by Google. Furthermore, the review of content from over 10,000 Turkish news websites revealed journalistic challenges within independent media, too.

Methodology

  • Scraping and analyis of 1.000.000+ recommendations for Turkish news on Google's search engine and Google News
  • Collection and analysis of 770.000 articles from 10.000+ Turkish news websites
  • Emotional analyis in 90.000 political articles and 1.300.000 million headlines, using GPT models
  • Thematic analysis of the coverage through 60 keywords

Key findings

  • Algorithmic preferences: a pronounced bias in Google's search results towards pro-government media outlets (74,8% recommendations)
  • News diversity: a concerning contraction in the diversity of Google's sources, with large-scale corporate media receiving preferential treatment (22 news websites generated 50+% of recommendations)
  • Negative coverage: predominant negative sentiments towards opposition political parties and figures

Full report: English

Tags: Turkey Business model Media pluralism Media ownership

The content of this article can be used according to the terms of Creative Commons: Attribution-NonCommercial 4.0 International (CC BY-NC 4.0) . To do so use the the wording "this article was originally published on the Resource Centre on Media Freedom in Europe" including a direct active link to the original article page.