TikTok and X algorithms showed far-right bias in Germany ahead of 2025 federal election
Research by Global Witness found that TikTok, X, and Instagram's recommendation algorithms showed substantial bias toward far-right political content in Germany ahead of the February 23, 2025 federal election, with TikTok's feed recommending far-right AfD party content at rates far exceeding the party's actual polling support.
What happened
Global Witness, a non-governmental organization, analyzed algorithmically sorted 'For You' feeds on TikTok, X, and Instagram to test for political bias ahead of Germany's federal election on February 23, 2025. Test accounts were set up to follow the four biggest German political parties equally, presenting as non-partisan users interested in political content.
TikTok's 'For You' feed showed 78% of recommended political content from unfollowed accounts supporting the far-right AfD party, while X showed 64% such support and Instagram showed 59%. On general left versus right-leaning bias, TikTok displayed right-leaning content 74% of the time, X showed 72%, and Instagram showed 59%.
Reported impact
The AfD party attracted approximately 20% support in polling, while receiving 78% of TikTok's algorithmic recommendations and 64% of X's recommendations. Non-partisan social media users in Germany were being exposed to right-leaning content more than twice as much as left-leaning content ahead of the federal election.
Index analysis
Global Witness noted a lack of transparency in how these algorithms work. Ellen Judson, a senior campaigner at Global Witness, suggested the bias might be an unintended side effect of algorithms maximizing engagement, though the right-wing skew across multiple platforms raises questions about whether additional factors were at play.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.