Report September 2025
Submitted
Democracy Reporting International's (DRI) Digital Democracy Programme Unit focuses on identifying trends in online discourse and potential threats to the integrity of information during political events and electoral periods across Europe and beyond. Our Digital Democracy team monitors and audits key elements of the online information space (social media platforms, consumer AI products) and formulates policy recommendations for various stakeholders in the technology and society ecosystem, including lawmakers, tech platforms, and civil society organisations. During the reporting period, DRI identified multiple threats to electoral integrity across Germany, Romania, and Poland, including impersonation by TikTok murky accounts, misinformation by chatbots, potential algorithmic bias in recommender systems, and reductions in platform commitments under the EU Code of Practice transitioning into the legally binding Code of Conduct. These threats risked misleading voters, amplifying extremist content, and undermining transparency and accountability. In response, DRI implemented a multi-pronged mitigation strategy, combining research, stakeholder engagement, advocacy, legal action, and public awareness efforts to strengthen compliance with EU digital regulations and safeguard civic discourse.
Key Findings and Actions during the Reporting Period:
Key Findings and Actions during the Reporting Period:
- TikTok Murky Accounts Research: DRI identified 606 inauthentic TikTok accounts across Germany, Romania, and Poland. We reported them through the Election Rapid Response System. Following an internal review based on its Terms of Services and Community Guidelines, TikTok took action to remove 414 of them. These accounts impersonated politicians or party pages, disproportionately promoting right-leaning candidates and parties. In Poland, accounts supporting Konfederacja generated 12 times more engagement than the second-most engaged party, highlighting the potential impact on voter perception and discourse, as well as some gaps in the understanding and regulation of political propaganda/advertisement.
- Chatbots and Generative AI: Monitoring of six major chatbots revealed persistent misinformation on electoral topics, despite some improvements aligned with DRI and EU Commission guidelines. In addition, far-right parties in Germany and Poland used AI-generated images and videos, often without disclosure, to misinform, reinforce narratives, and increase engagement, demonstrating the growing role of generative AI in campaigns.
- Recommender Systems and Algorithmic Bias: In the context of the 2025 German federal election, we explored the recommender algorithms on TikTok and Instagram to assess how users’ political interest and preferences shaped the amount of political content they encountered, and how these dynamics varied across both platforms. We also analysed whether recommended content aligned or not with users’ pre-defined political positions. To ground this analysis, we conducted a literature review on political exposure bias during the 2024 U.S. and 2025 German elections. Results show that platforms amplified right-leaning or far-right content even for neutral users, most strongly on X and TikTok, and weakly on Instagram. Users aligned with far-right parties received more targeted recommendations, and less cross-party recommendations. This persistent bias potentially poses systemic risks, including polarisation and radicalisation of online political discourse.
- Platform Rollbacks under the CoC: Major platforms scaled back commitments under the EU Code of Practice transitioning to the Code of Conduct, reducing fact-checking, political advertising transparency, and researcher support by 31%. Microsoft, Google, and TikTok withdrew key measures, and all platforms abandoned Commitment 27 on researcher data access, raising concerns about the CoC’s effectiveness in ensuring accountability and combating disinformation.
- Legal Action and Advocacy: DRI sued X under DSA Article 40(12), testing the platform’s obligation to provide timely access to publicly available data ahead the German federal elections in February 2025. The Court’s ruling in May gave extensive interpretation of the data access provision and declared itself competent to review the case, an important precedent that CSOs and researchers can sue in all EU Member States (and not only in Ireland at X HQ). A public webinar unpacked the Berlin court ruling and its implications, providing guidance for other CSOs seeking to pursue similar advocacy.