Hogan Lovells publishes report on the impact of Generative AI on the Right to Free and Fair Elections
Press releases
| 20
January
2025
London, 20 January 2025 - Global law firm Hogan Lovells has today published a report examining whether there are positive legal obligations on States that are party to the European Convention on Human Rights (“ECHR”) to intervene to protect the integrity of their electoral systems from threats hypothesised to emerge from the misuse of generative AI.
The advent of generative AI tools brings with it the risk that deepfakes (artificial video and audio content that is nearly indistinguishable from live recordings) will be used to influence what people believe about political candidates, political issues, and voting locations and times. Commentators have expressed concern that deepfakes could, in turn, be used to sway individuals’ voting choices or prevent them from voting altogether.
The report is jointly authored by Hogan Lovells’ Public Law & Policy and Pro Bono teams, led by Partner Charles Brasted with contributions including from London associates Fraser Eccles and Haylea Campbell, trainees Cara Nicholson and Elmina Marriot and guidance from International Pro Bono Partner Yasmin Waljee OBE. The team’s research was conducted in collaboration with Demos, a leading cross-party think tank, as part of their strategic focus area on ‘Trustworthy Technology’.
Key findings from the report include:
- 1) At this time, the obligations imposed on states under Article 3 Protocol 1 of the ECHR (the Right to Free and Fair Elections) have not been found to include positive obligations to protect voters from the impacts of mis/disinformation facilitated by generative AI. The ECHR is, however, a ‘living document’, and so the law could develop in this direction in the future. Some previous judgments of the European Court, addressing circumstances that are arguably analogous, could provide a potential basis for future strategic litigation in this space.
- 2) Even though it may not create a positive obligation to do so, the ECHR does not prevent, in principle, states from taking action to address the impacts of dis/misinformation enabled generative AI. Any such intervention, however, must not endanger other ECHR rights – including the right to freedom of expression.
- 3) If a positive obligation to safeguard elections from the impacts of generative AI was found to exist, an applicant would still need to convince the Court that that positive obligation had in fact been violated by a contracting state, which may be challenging.
Commenting, partner Charles Brasted, said: “It has been a pleasure to collaborate with Demos on this timely report. Following a year in which almost half the world’s population went to the polls, and as the development of AI technology has continued to accelerate, this is a good opportunity to take stock of the law in this space and where it might develop in the future”.
The report is available
here.