Searching for Trouble: Problematic Content and Political Advertising on Search Engines During German 2025 Federal Election

The early federal election in Germany, scheduled for February 2025, comes at a time of significant political turmoil and uncertainty. Triggered by the collapse of the ruling coalition, this election is taking place seven months ahead of schedule. Amid the decline in support for the coalition parties, populist far-right Alternative for Germany (AfD) and left Sahra Wagenknecht Alliance (BSW) have been gaining traction among voters. In this climate of political instability, digital platforms are expected to play a significant role in shaping voter opinions and impacting the outcome of the election. Platforms are likely to be central to the dissemination of campaign messaging by all parties, but their extensive use by populist actors raises concerns over misinformation and targeted manipulations, which can contribute to democratic erosion. Search engines, in particular, are important platforms to investigate for the potential presence of biases in favor or against specific candidates and the possible amplification of misinformation and negative campaigning via political microtargeting. Considering that voters often use search engines to inform themselves about elections the quality of information environments on search engines is integral for ensuring democratic standards during crucial political events. Moreover, search engines remain one of the main platforms that citizens use to seek political information and tend to trust it (Dutton et al., 2017). However, studies have demonstrated that search engines can amplify stereotypes and potentially contribute to discrimination, in particular of vulnerable groups (Urman & Makhortykh, 2022). Manipulation of platform algorithms by foreign governments or domestic political actors has been another area of concern. Some of these manipulations fall in the realm of computational propaganda, where platforms’ algorithms are used to facilitate and amplify the spread of false content (Bradshaw, 2019). Considering evidence that information exposure on search engines can affect the voting process (Epstein & Robertson, 2015), the distribution of misinformation and targeted attacks (e.g. via poltical ads) in search results has the potential to destabilize electoral processes. To investigate how grounded these concerns are in the case of German federal elections, we performed a virtual agent-based algorithmic audit to study the presence and distribution of problematic content in organic results and political advertising on two most commonly used search engines in Germany, Google and Bing, prior to the elections. As search queries, we used names of political parties and key federal-level politicians together with a selection of issue-specific queries based on the Wahl-o-Mat voting advice platform; furthermore, we integrated a set of politics-related queries from Google Trends for Germany from the run-up to the election. To investigate how different topics are prone to the emergence of problematic content in search results, we use a combination of content analysis to identify the presence of hyperpartisan, alternative, and conspiratorial (HAC) media in search results and discourse analysis to analyze the content of political ads included in search results.