EU AI Act mandates labeling of chatbots and deepfakes

A new set of European Union transparency rules governing artificial intelligence has taken effect, requiring clear disclosures for chatbots, deepfakes, and other AI-generated content. The obligations, part of the bloc's landmark Artificial Intelligence Act, became enforceable on August 2, 2026, granting authorities new powers to inspect systems and levy substantial fines for non-compliance.

The core requirements mandate that providers of AI systems designed to interact directly with people, such as chatbots, must inform users they are communicating with artificial intelligence. Separately, deployers of systems that generate or manipulate image, audio, or video content to create realistic depictions of people, objects, or events-commonly called deepfakes-must disclose that the content has been artificially produced. For AI systems that generate or manipulate content, providers are also required to embed machine-readable markers in the outputs to facilitate detection, where technically feasible.

The rules also cover other specific applications. Deployers using emotion recognition or biometric categorization systems must notify individuals exposed to that technology. A separate...