Is it always clear that we are interacting with AI or consuming AI-generated content? More and more, the answer to this question is no? This matter is adressed by Article 50 of the European Union’s Artificial Intelligence Act (AI Act) wich becomes applicable on 2 August 2026 (Artificial Intelligence Act, 2024, Art. 50, 113).
Article 50 does not simply introduce the same warning for every type of image, text, audio, or conversation supported by AI (Artificial Intelligence Act, 2024, Art. 50). Rather, it establishes specific obligations for providers and deployers of AI systems. It therefore turns transparency from a matter of voluntary practice into a legal duty in defined circumstances.
Scope of Article 50 of the EU AI Act
The EU AI Act makes a clear distinction in Article 3 between providers and deployers. A provider develops an AI system or has an AI system developed, and places it on the market or puts it into service under its own name or trademark. A deployer uses an AI system under its authority, except where the use is personal and non-professional (Artificial Intelligence Act, 2024, Art. 3). This distinction between provider and deployer determines who must implement technical transparency measures, and who must disclose relevant information to persons exposed to particular AI systems or content.
AI-Interaction and Synthetic Content
The first obligation laid out in Article 50 pertains to direct AI-human interaction (Artificial Intelligence Act, 2024, Art. 50). A provider of an AI system that communicates directly with natural persons must design the system in such a way that the natural person is informed that they are interacting with an AI. This is unless it is already clear from the context and circumstances of use. Such a disclosure would have to be “at the latest at the time of the first interaction or exposure” (Artificial Intelligence Act, 2024, Art. 50). Should the disclosure be only somewhere in the general terms and conditions, it would not reach the central purpose of the regulation.
Article 50(2) dictates that AI systems that generate synthetic audio, images, videos, and text must ensure that their outputs are marked in a machine-readable format and detectable as artificially generated or manipulated (Artificial Intelligence Act, 2024, Art. 50(2)). This marking must be effective, interoperable, reliable and robust insofar as technically feasible. The provision recognizes that there are certain limits to visible labels. Those could, for example, be removed or covered. Thus, machine-readable information provides a safer method to ensure that AI content is identifiable by platforms, moderators and detection systems. Also, the Commission’s Code of Practice on Transparency of AI-Generated Content suggests practical implementation measures. This might include content marking and detection tools(European Commission [EC], 2026). While the Commission’s Code of Practice on Transparency of AI-Generated Content is voluntary, the duties layed out in Article 50 are binding.
Article 50(3) further requires deployers of emotion-recognition or biometric-categorisation systems to inform natural persons exposed to those systems about their operation (Artificial Intelligence Act, 2024, Art. 50(3)). It is important to note that this obligation is separate from the duties concerning AI-human interaction, synthetic-content marking and deepfakes.
Deepfakes
Under Article 50(4), deployers are required to disclose when an AI-generated or manipulated image, audio, or video content constitutes a deepfake (Artificial Intelligence Act, 2024, Art. 50(4)). The purpose of this is to prevent persons from mistaking apparently authentic deepfakes for authentic material or genuine evidence. This provision is also applicable to AI-generated or manipulated text published with the purpose of informing the public about matters of public interest. According to European Commission (n.d.), this does not apply to text that underwent significant review or editorial control by a human, where a natural or legal person assumes editorial responsibility. Any publisher of AI-generated or AI-manipulated text that is intended to inform the public falls under this rule, including but not limited to traditional media.
Application and Enforcement
In general, Article 50 applies from 2 August 2026. However, providers of systems placed on the market before that date have until the 2nd of December 2026 to comply with the Article 50(2) machine-readable marking obligation (European Commission, n.d.).
Under Article 99(4), Member States must provide for penalties for Article 50 infringements. The maximum fine may be up to €15 million or, for an undertaking, up to 3% of worldwide annual turnover for the preceding financial year, whichever is higher (Artificial Intelligence Act, 2024, Art. 99(4)).
Conclusion
Article 50 is unlikely to stop the development and use of AI. Instead, it requires transparency where persons may otherwise be unaware that they are interacting with AI or encountering synthetic or manipulated content.
The main requirements are clear: AI interaction must be disclosed, the detection of synthetic content must be enabled, deepfakes must be labelled, persons exposed to emotion-recognition and biometric-categorisation systems must be informed, and certain AI-generated or manipulated public-interest text must be disclosed unless it has undergone human review or editorial control and an identifiable person or entity assumes editorial responsibility.
For organisations, the practical response is to identify whether they act as a provider or deployer, document relevant AI uses, implement appropriate marking and disclosure measures, and allocate responsibility for AI-assisted public communications.
References
European Commission. (n.d.). Frequently asked questions. AI Act Service Desk. Retrieved August 31, 2026, from https://ai-act-service-desk.ec.europa.eu/en/faq
European Commission. (2026, June 10). Code of practice on transparency of AI-generated content. Shaping Europe’s Digital Future. https://digital-strategy.ec.europa.eu/en/policies/code-practice-ai-generated-content
European Parliament and Council of the European Union. (2024). Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonised rules on artificial intelligence and amending Regulations (EC) No 300/2008, (EU) No 167/2013, (EU) No 168/2013, (EU) 2018/858, (EU) 2018/1139 and (EU) 2019/2144 and Directives 2014/90/EU, (EU) 2016/797 and (EU) 2020/1828 (Artificial Intelligence Act). Official Journal of the European Union, L 2024/1689. http://data.europa.eu/eli/reg/2024/1689/oj
