AI-generated deepfakes pose growing threats to elections globally, prompting varied legal responses across Europe.
🔗 Read the article on wyborcza.pl
[This is an automatically generated summary, for reference only]
A historical example from the 1960s showed how fabricated compromising images and false reports of protests could influence elections through television broadcasts. Today, this is a reality, with instances of deepfakes and AI-generated content appearing during Polish presidential election campaigns in 2025 to discredit candidates. Furthermore, generated media are used post-election to allege electoral fraud. Research from 2025 indicated that brief interactions with large language model chatbots could shift voter preferences through microtargeting. Russia is also noted for using AI to create anti-Ukrainian fake footage.
Across the EU, there has been a significant rise in deepfake incidents, particularly in Bulgaria, Portugal, and Belgium. The EU’s legal response involves three pillars: the AI Act, the Transparency of Political Advertising Regulation (TTPA), and the Digital Services Act (DSA), which impose heavy fines for non-compliance. However, current legislation is deemed insufficient as individual member states lack adequate powers to quickly identify and remove high-quality deepfakes.
In Poland, no specific national law regulating AI in elections exists; a task force has been announced by the Ministry of Justice to address deepfakes. The President of the Personal Data Protection Office noted that existing laws are often not applied rigorously, with new forms of violations sometimes being dismissed as jokes despite serious potential consequences, advocating for regulations similar to those in France, Italy, or the US that explicitly criminalize image manipulation via algorithms and mandate swift removal procedures.
France has some of Europe’s strictest rules, including a 2024 law imposing penalties for publishing deepfakes without consent unless clearly labeled as manipulated content, allowing political actors to request immediate takedown orders before elections. Lithuania has fined a politician for using AI-generated Russian dubbing in parliamentary broadcasts and is actively updating its AI strategy to prioritize cyber resilience against deepfakes used for disinformation. Italy’s digital services law focuses on platform moderation rather than the intentional political use of synthetic media, leading some news outlets to propose voluntary commitments from political parties regarding AI use in campaigns, while a legislative proposal aims to ban misleading AI-generated content during elections. The overall concern is that sophisticated deepfakes are becoming indistinguishable from reality and can be used by politicians to undermine authentic evidence by claiming it is AI-generated.
