The key clash during the European Parliament's joint public hearing on "Democracy and Elections in the AI Era" on 18 March 2026 centered on the nature of the threat AI poses to electoral integrity, with prominent panelists diverging sharply over the necessity of new legal frameworks versus enhanced enforcement of existing rules, and the balance between national sovereignty and EU-level governance.
Sofia Calabrese from the European Partnership for Democracy and Ines Narciso from the SANS Institute took the side emphasizing that the EU already possesses a robust legal framework encompassing the Digital Services Act (DSA) and the AI Act. Calabrese focused on enforcement failures, urging stronger risk assessments of very large platforms and better use of political advertising regulations. Narciso pointed to operational capacity shortfalls, interoperability issues, and platform design flaws rather than regulatory gaps as the main defects needing attention. Contrarily, Juliane Muller of International IDEA adopted a more moderate stance, advocating for building upon existing legislation but closing gaps to cover AI practices impacting election administrative systems.
Another divisive theme was the tension between national sovereignty and EU-wide responses to AI-related disinformation and manipulation. András László (PfE) insisted that national elections remain within member states’ exclusive competence and cautioned against EU overreach possibly stifling democratic debate under the guise of fighting interference. By contrast, Muller and other speakers stressed that fragmented national strategies invite cross-border manipulation, advocating for EU-wide coordination mechanisms to bolster resilience while respecting administration kept at national levels.
This hearing, held by the European Parliament committees LIBE, AFCO, and the Special Committee on the European Democracy Shield, with participation from AFET and IMCO, assembled expert insights and MEP interventions highlighting issues such as deepfakes, algorithmic amplification, bot identification, provenance labeling, and more.
Concrete proposals included calls for interoperable EU data collection, pre-election stress testing (Narciso), stronger DSA enforcement pushing back against platforms and ensuring data access for independent risk analysis (Calabrese), plus enhanced operational capacity, including more expertise in the AI Office and a better-resourced European Centre for Democratic Resilience (Muller). These initiatives aim to increase transparency and authenticity signals, introduce bot identification, and enforce risk assessments for AI systems influencing elections.
In contrast, some speakers such as Christine Anderson (ESN) warned against conflating normal democratic discourse with manipulative AI use and called for strict limitations on governance powers. Gheorghe Piperea (ECR) raised concerns about responsibility assignment with biased AI outputs possibly overwhelming legal checks.
increasing EU-level supervisory powers and operational staff against respecting member states’ electoral sovereignty; enhancing transparency and bot exposure versus avoiding over-broad content removal; and prioritizing enforcement with concrete operational deadlines against new legislative expansion.
Stakeholders impacted by these discussions are diverse. Political platforms and media companies face heightened compliance and transparency demands, entailing moderate increased operational costs. Electoral authorities across member states would gain from enhanced expertise, coordination, and resources but must adapt to new procedural tasks. EU civil society and voters, targeted by disinformation and manipulated content, may benefit from improved transparency, authenticity markings, and citizen digital literacy efforts, strengthening democratic resilience. Finally, national governments confront the trade-off between preserving sovereignty over elections and cooperating at EU level to counter transnational AI threats.
Going forward, the Parliament’s ongoing debate on digital democracy will likely focus on reconciling these cleavages through legislative amendments to the AI Act and DSA, combined with increased funding for operational preparedness and platform supervision. It remains uncertain how quickly member states will agree on these proposals, given differing views on sovereignty and democratic autonomy. However, the hearing underscores an EU-wide consensus that AI profoundly challenges election integrity, demanding both legal and practical responses that navigate a complex regulatory and political landscape.