What is Article 50 of the EU AI Act?
The AI Act (Regulation (EU) 2024/1689) came into force on 1 August 2024, and the transparency obligations under Article 50 apply from 2 August 2026.
Although Article 50 is an EU requirement, its reach extends beyond businesses established within the EU. It also applies to UK organisations that:
- place AI systems on the EU market, or
- whose AI system outputs are used in the EU.
From the Commission’s FAQ on Article 50:
People must be notified when they are interacting with an AI system from the start of the first interaction in a clear and distinguishable manner and in accordance with accessibility requirements.
This involves disclosures for AI systems such as chatbots, AI companions, AI agents, deepfakes, avatars, emotion recognition systems, biometric categorisation tools, and AI-generated text meant to inform the public and which has not undergone human review or editorial control.
In this blog, we focus on key Article 50 compliance actions for UK companies.
The four key transparency obligations under Article 50
Article 50 is designed to ensure people know when they are interacting with AI or viewing AI-generated content. It focuses on transparency, helping individuals make informed decisions when they encounter information produced by AI systems.
Article 50 places obligations on providers and deployers of certain AI systems:
- Providers must design AI systems intended to interact directly with individuals so that individuals are informed when they are interacting with those AI systems, unless this is obvious to a reasonably well-informed individual.
- Providers of AI systems generating or manipulating synthetic audio, image, video or text content must mark outputs and make detectable as artificially generated or manipulated.
- Deployers of AI systems must inform individuals when they are exposed to emotion recognition and biometric categorisation systems.
- Deployers of AI systems must disclose some content – deepfakes or text publications on matters of public interest without human review or editorial control – as artificially generated or manipulated.
What are the penalties for breaching Article 50?
Fines for companies breaching these transparency obligations can reach up to 3% of worldwide annual turnover in the preceding financial year or EUR 15 million, whichever is higher.
Oversight is by:
- the national market surveillance authorities;
- the Commission’s AI Office (for systems under its supervision); and
- the European Data Protection Supervisor (when EU institutions are providers / deployers).
Proportionality applies to SMEs and SMCs.
Beyond potential financial penalties, organisations should also consider the broader business implications of non-compliance. Transparency around AI use is increasingly linked to customer trust, brand reputation and responsible AI governance, particularly in sectors where organisations rely on AI-generated information or customer interactions.
Official guidance and resources on Article 50 compliance
The European Commission has published:
- a 38-page voluntary Code of Practice on Transparency of AI-generated Content (10 June 2026),
- complemented by its 50-page non-binding Guidelines on transparency obligations for providers and deployers of AI systems (20 July 2026).
A stated aim of the code is ‘to promote the uptake of human-centric and trustworthy artificial intelligence’.
In addition, there are other resources, including:
- Quick Facts,
- FAQs, and
- a set of EU icons.
As of 31 July 2026 (ahead of the application of legal obligations), there were 190 signatories to the Code of Practice.
On 30 September 2026, over 90 signatories of the code established two task forces to support the implementation of the code.
Article 50 obligations and responsibilities
| Within Article 50 | AI system | obligation | responsibility |
| Article 50(1) | System interacts directly with people | Informing | Providers |
| Article 50(2) | System generates synthetic content (e.g., audio, image, video, text) | Marking and detection | Providers |
| Article 50(3) | An emotion-recognition system / biometric categorisation system | Informing | Deployers |
| Article 50(4) | System generates or manipulates audio, image or video to create deep fakes / text to inform public | Disclosing | Deployers |
| Article 50(5) | Across requirements in Articles 50(1) – (4) | Information provided is clear, distinguishable, at first interaction / exposure latest, accessible | Providers and deployers |
For many organisations, one of the first challenges is determining whether they are acting as a provider, a deployer, or both.
Knowing the distinction is important because different transparency obligations apply depending on the role performed.
Distinguishing providers and deployers
Whilst a provider is broadly the developer of the AI system and the deployer the party using the AI system in a professional capacity, we have set out below extracts from the Commission’s FAQs which distinguish a provider from a deployer:
Who is a provider of an AI system and what transparency obligations apply to them?
According to Article 3(3) of the AI Act (https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-3), providers of AI systems are natural or legal persons, public authorities, agencies or other bodies that develop AI systems, or have them developed, and place them on the EU market or put them into service under their own name or trademark. This is irrespective of whether those providers are established or located within the EU or in a third country.
Providers of AI systems established or located outside the EU are also subject to the provisions of the AI Act if the output of their AI system is used in the EU…
Who is a deployer of an AI system and what transparency obligations apply to them?
Deployers of AI systems are natural or legal persons, public authorities, agencies or other bodies using AI systems under their authority, excluding use for personal, non-professional activities…
Transitional compliance period for AI systems on the EU market before 2 August 2026
There is a limited grace period, introduced by the Digital Omnibus on AI (Regulation (EU) 2026/1744), until 2 December 2026 for legacy generative AI systems on the market before 2 August 2026 for marking and detection sub-obligations.
Recital 38 reads:
To allow sufficient time for providers of generative AI systems subject to the marking obligations laid down in Article 50(2) of Regulation (EU) 2024/1689 to adapt their practices within a reasonable time without disrupting the market, it is appropriate to introduce a transitional period of four months for providers who have already placed their systems on the market before the 2 August 2026.
Importantly, this grace period does not apply to AI systems placed on the market on or after 2 August 2026, which must comply with the transparency requirements under Article 50 where they apply.
Ten practical steps towards Article 50 compliance
We conclude with words from the European Commission’s overview to the Guidelines, alongside ten practical steps organisations can take to assess whether Article 50 applies and begin preparing for compliance.
These obligations are intended to foster trust and integrity in the information ecosystem. People should know when they are interacting with AI or exposed to AI-generated content. This will help them make informed decisions, calibrate their trust and reliance on AI and avoid mis information or deception.
- Determine whether your AI system is placed on the EU market, or whether its output is used or intended to be used in the EU.
- Determine whether you act as a provider, deployer, or both.
- Identify AI interactions and content generation / manipulation that trigger Article 50 transparency.
- Notify users from the first interaction or exposure in a clear, distinguishable, and accessible manner.
- For providers, apply machine-readable marks to synthetic content generated or manipulated, unless an exemption applies such as where the system only assists standard editing.
- For deployers, disclose emotion recognition and biometric categorisation where required.
- For deployers, clearly label deepfakes and AI-generated text on matters of public interest where human review or editorial control has not taken place
- Monitor key compliance deadlines, including from the application date of 2 August 2026; existing generative systems on the market before 2 August 2026 must meet marking and detection obligations by 2 December 2026.
- Keep records to demonstrate compliance.
- Use the Commission’s Code of Practice, Guidelines, FAQs, and EU icons to guide implementation.
Article 50 does not explicitly apply to general-purpose AI (GPAI) models, although rules on GPAI may interact with Article 50 (see, for example, para 2.6 of the Guidelines: Interplay with rules applicable to general-purpose AI models and/or systems).
How we can help
Determining whether your organisation is acting as a provider, a deployer, or both is not always straightforward, particularly where multiple AI tools, third-party systems or cross-border operations are involved.
As organisations increasingly adopt AI across customer services, content creation, HR, compliance and operational functions, understanding applicable transparency obligations is becoming an important part of broader AI governance and risk management.
Our experienced team of technology solicitors can support you if you would like to learn more about how Article 50 interplays with other provisions within the EU’s AI Act and the wider regulatory landscape.
This content is provided for general informational purposes only and does not constitute legal advice. It is not intended to address the circumstances of any individual or entity, nor should it be relied upon as a substitute for specific advice from a qualified solicitor. The information reflects the legal position as at the date specified and may be subject to change. If you require advice on a specific matter, please contact us directly.





