New transparency obligations under the European Union’s AI Act took effect in August 2026, creating disclosure requirements for chatbots, deepfakes and certain artificial-intelligence content used to inform the public.
What the rules require
The European Commission says people should be informed when they are interacting with a machine where that knowledge is necessary for an informed decision. Providers of generative AI must also make AI-generated material identifiable, while certain deepfakes and AI-generated text on matters of public interest require clear labelling.
The measures form one part of the EU’s risk-based AI framework. Systems considered minimal risk generally face no new obligations, while prohibited practices and rules for general-purpose AI began applying earlier.
Enforcement enters a new phase
From August 2, 2026, the European AI Office and national authorities became responsible for implementation and enforcement. The Commission says the AI Office can request documentation, evaluate general-purpose models, require corrective action and impose fines where the law permits.
Some high-risk-system deadlines extend into 2027 or 2028 following later amendments. That staggered calendar means companies must identify which provisions apply now rather than treating the Act as a single compliance date.
Impact on platforms and publishers
Developers need technical methods to identify generated content, while publishers and platforms must decide when disclosures are sufficiently clear. The rules are particularly relevant to synthetic audio, video and text that could be mistaken for authentic reporting or political communication.
The framework is likely to influence product design outside Europe because global services may adopt common disclosure systems. Practical enforcement, technical standards and court interpretation will determine how consistently the rules work.
