EU AI Act (2024 – 2025) & Elections

The first continent-wide AI law forces every large platform (TikTok, Instagram, YouTube, Facebook, X) to show its workings when algorithms shape political speech. It is designed to kick-in before the 2025 European Parliament elections.

Main obligations for TikTok & Co.

• Risk Audits & Mitigation
  Annual risk reports on algorithmic manipulation, political deep-fakes and foreign influence operations.
• Transparent Recommenders
  A clear label when For-You feeds are driven by profiling; users must get a “no-fuzz chronological” option.
• AI Content Labelling
  All synthetic or AI-generated audio-visual posts must carry a persistent on-screen watermark and metadata tag.
• Easy Flag + Human Review
  Citizens can flag suspected deep-fakes or coordinated disinformation — a human moderator must review within 24 h during election periods.
• Public Ad Libraries
  Real-time searchable database of every paid political or “issue-based” ad, including sponsor & spend.

➜ Read the full text at the EU Official Journal  |  ➜ Commission overview: digital-strategy.ec.europa.eu