The European Commission launched a public consultation on Tuesday 10 November to gather stakeholders’ views, until Monday 23 November, on draft guidelines updating the measures that very large online platforms (VLOPs) should adopt to address the risks that generative artificial intelligence (AI) may pose to electoral processes.
The update proposes five main changes:
(1) stronger protection against foreign information manipulation and interference (FIMI): rapid detection of coordinated inauthentic behaviour and dismantling of the networks behind it;
(2) new requirements for platforms’ ‘recommender systems’: assess suspicious trends before they go viral, test their algorithms for manipulation, and keep a record of changes made to recommender systems during the six months before an election;
(3) stronger protection against AI-generated content: ‘deepfakes’ should be clearly labelled, and the label should remain when they are shared;
(4) better protection for candidates: dedicated channels on platforms and swift handling of threats, harassment and impersonation;
(5) greater accountability: platforms should report to the Commission and national authorities, within 30 days of an election, on how they mitigated risks and publish a public version of the report.
See the draft guidelines: https://aeur.eu/f/ocb
Take part in the public consultation: https://aeur.eu/f/oca (Original version in French by Mathieu Bion)