Article 50 of the AI Act on labeling AI content entered into force; ČTÚ alerted companies to new obligations

The Czech Telecommunication Office (ČTÚ) alerted companies and organizations in the Czech Republic to the entry into force of Article 50 of the EU Artificial Intelligence Act (AI Act), which took place on 2 August 2026. The provision introduces transparency obligations for providers of AI systems (entities placing a system on the market or putting it into service) and for deployers (persons using an AI system for activities other than personal, non-professional activities). According to ČTÚ, correctly identifying one's own role is essential because some obligations apply only to providers, others only to deployers, and anyone using their own AI system must comply with both sets of obligations.

Under paragraph 1, providers must ensure that individuals are informed that they are communicating with a chatbot, voicebot, or another system intended for direct interaction with people. Under paragraph 2, all audio, image, video, or text output from AI systems must be marked in a machine-readable and detectable format, with exceptions specified by law. Under paragraph 4, deployers are required to label deepfakes and AI-generated or modified text published for the purpose of informing the public about matters of public interest if it may give the impression of authenticity; ČTÚ recommends clear labels such as “AI generated” or “AI modified” for this purpose, in accordance with the Code of Practice on Transparency of AI-Generated Content. Under paragraph 3, deployers using AI for emotion recognition or biometric categorization must inform the persons concerned.

Compliance with these obligations in the Czech Republic is overseen by, among others, ČTÚ as a market surveillance authority, and in some cases by the European AI Office (AI Office) instead, depending on the role of the entity and, where applicable, on the provider of the general-purpose model on which the system is based. ČTÚ recommends that companies assess their obligations in good time and take measures to comply with the transparency rules. Government Commissioner for AI Lukáš Kačena added that the rules do not apply to end users who use AI for entertainment and non-commercial purposes, that the AI Act, as a regulation, applies directly without the need for national legislation (although certain details, including confirmation of the role of ČTÚ as the main supervisory authority, will be addressed by the adaptation law being prepared), and that violations carry potential fines of up to 360 million Kč or 3 % of worldwide turnover, with exceptions for small and medium-sized enterprises.