On 12 July, the European Parliament published the AI Act. The EBU welcomes this step that establishes a binding and comprehensive AI rulebook for Europe. 

Wouter Gekiere, Head of the Brussels Office, EBU, said: “AI offers significant opportunities, but it also introduces new risks with implications for the media. These include an increase in dis- and mis-information, the proliferation of low-quality content, threats to media diversity, and challenges to existing copyright laws. The AI Act seeks to address some of these issues with new obligations to identify AI systems that interact with individuals and to label content generated or altered by AI. The legislation also requires general-purpose AI providers to disclose the data used to train their models, allowing content providers, such as public service media, to gain greater transparency on how AI systems use their content.” 

EBU data shows that most public service media are already using AI enabled tools to enhance the delivery of quality content and services to audiences. With the AI Act now in place, Public Service Media are ready to support EU institutions to ensure the effective implementation of the new regulations. 

Next steps: 

• 1 August 2024: the AI Act enters into force 
• February 2025: rules on prohibited AI systems start applying 
• May 2025: codes of practice (e.g. for General Purpose AI models) must be ready 
• August 2025: rules on General Purpose AI models start applying 
• August 2026: the whole AI Act applies – except for Article 6(1) (obligations for high-risk systems defined in Annex II) 
• August 2027: Article 6(1) & corresponding obligations apply