This briefing paper explores the impact of AI-enabled disinformation and fake content on democratic processes, and investigates the potential obligations of states to protect elections from mis- and disinformation. In particular, the paper considers whether AI-generated content poses a threat to the right to free and fair elections as articulated by the European Convention on Human Rights (ECHR).

The paper discusses how AI, particularly deepfakes, can affect elections by spreading false information, manipulating voters, and intimidating candidates. It includes a legal analysis of the ECHR, focusing on states' obligations to ensure free and fair elections and freedom of expression. There is no specific case law obligating states to prevent AI undermining elections, but notes that the ECHR is a ‘living document’ that could develop in this direction.

Recommendations

1. Reduce instances of deceptive AI-enabled disinformation by 
   1. establishing responsible standards for the use of generative AI by political parties during election campaigns,
   2. requiring social media platforms to label artificially generated content,
   3. clarifing and granting enforcement powers to regulatory bodies, and
   4. enabling the Electoral Commission to notify the public about false or misleading content.  
2. Strengthen societal resilience to AI-enabled mis- and dis-information
3. Defend access to polling
4. Defend the ability of candidates to freely stand for election by 
   1. criminalising the use of deepfakes to produce and distribute non-consensual pornographic content, 
   2. ensuring that victims of online abuse are able to claim compensation, and
   3. requiring search engine providers to exclude and/or downrank AI generated non-consensual imagery.