EU AI Act High-Risk Deadline Pushed to 2027, But Transparency Rules Now in Force
The EU extended its deadline for high-risk AI system compliance by sixteen months, with standalone systems now due by 2 December 2027 and product-embedded systems by 2 August 2028. The shift was formalised through Regulation (EU) 2026/1744, which entered into force on 27 July 2026 — just six days before the original deadline. The postponement was not a policy retreat but a practical one: harmonised technical standards and accredited certification bodies required to demonstrate compliance were not yet in place. However, transparency obligations under Article 50 took effect as scheduled on 2 August 2026, requiring AI systems to identify themselves to users, label synthetic media in machine-readable formats, and disclose AI-generated public-interest content unless a named human editor holds accountability for it. The rules signal that the EU is not abandoning its framework but waiting for the certification infrastructure needed to enforce it.
This is an AI-generated summary. ShortSingh links to the original source for the complete article.
Discussion (0)
Log in to join the discussion and vote.
Log in