BG2.info

European Union: New transparency rules for the use of AI and deepfakes come into force

Artificial Intelligence · 2026-07-26 15:25:51

Patrick Gamelkoorn

New transparency requirements in the EU

As of August 2, new provisions under the Artificial Intelligence Act (AI Act) are coming into force in the European Union. These rules affect chatbots, generative artificial intelligence (AI), and deepfake technologies, aiming to provide greater clarity for users.

According to the new regulations, AI systems must mandatory inform people that they are interacting with a machine. Content that is created or significantly modified through artificial intelligence must also be accordingly labeled.

Labeling and deepfakes

Technology developers are required to implement machine-readable labeling in images, videos, audio recordings, and texts. Special attention is paid to deepfakes – if they realistically recreate people, their use must be disclosed. However, these requirements do not cover standard text editing, such as spell or grammar checks, nor content that is obviously fictional.

For most users, the changes will be barely noticeable. Visible signs of generated content remain voluntary, but technical markers and metadata will be mandatory. Generative services already on the market have a transition period until December 2, 2026, to implement the necessary labeling.

Sanctions and future restrictions

Control over the enforcement of the rules will be carried out by national regulatory authorities in EU member states. In case of violations of transparency requirements, companies can be fined up to 15 million euros or up to 3% of their annual global turnover.

You may also like

The regulation also provides for next steps: from December 2026, the dissemination of certain categories of dangerous AI-generated content will be prohibited, and during the 2027–2028 period, even stricter rules for high-risk systems will come into force.

Пълната версия на статията · © 2026 BG2.info