Authors: Patrick Massa, Martina Grech, Estelle Scicluna – WH Partners
Following the EU AI Act’s entry into force on 1 August 2024, its staggered implementation timeline marked 2 August 2026 as a critical milestone for its general application. Whilst the Digital Omnibus on AI has considerably changed the significance of this date by delaying the application of various provisions, 2 August 2026 remains a key checkpoint in the AI Act’s application timeline.
2 August 2026 has primarily triggered the application of the transparency obligations under Article 50, together with the operation of the AI Act’s supervisory and enforcement provisions, most notably the enforcement provisions concerning general-purpose AI models.
Transparency Obligations
From 2 August 2026, Article 50 of the AI Act has become applicable to providers and deployers of certain AI systems. These transparency obligations are intended to mitigate the risks of AI misuse that may result in deception, manipulation, and the spread of misinformation.
The key transparency obligations are:
Informing individuals when they are interacting with AI: providers of AI systems designed for direct interaction with natural persons must ensure that such AI systems are developed and designed in a manner that clearly informs users that they are interacting with an AI system. This requirement may not apply where the use of AI is self-evident.
Labelling of AI-generated content: providers of AI systems that generate or manipulate images, videos, audio or text, must ensure that such outputs are marked in a machine-readable format and detectable as artificially generated or manipulated.
Transparency around emotion recognition and biometric categorisation systems: deployers of emotion recognition or biometric categorisation systems are required to inform individuals when they are exposed to the operation of such systems.
Disclosure of deepfakes: deployers of AI systems generating deepfakes must disclose that the content has been artificially generated or manipulated. Notably, the AI Act defines deepfakes as “AI-generated or manipulated image, audio or video content that resembles existing persons, objects, places, entities or events and would falsely appear to a person to be authentic or truthful”.
Disclosure of AI-generated content on matters of public interest: where an AI system is used to generate or manipulate text that is published by deployers to inform the public on matters of public interest, the content must be clearly disclosed as AI-generated. This disclosure is not required where the content has undergone human review and editorial approval before publication, or where a legal person assumes editorial responsibility for the published content.
Any disclosure required under Article 50 of the AI Act must be made no later than the user’s first interaction with the AI system. The information must be presented in a clear, distinguishable and easily understandable manner to ensure that individuals are adequately informed from the outset of their interaction with or exposure to the AI system.
To support compliance with the transparency requirements set out in Article 50 of the AI Act, the European Commission has recently adopted guidelines aimed at providers and deployers of AI systems subject to the transparency obligations. The guidelines clarify key concepts and definitions, and offer practical guidance on the measures that providers and deployers may implement to satisfy the relevant obligations.
The publication of these guidelines follow the European Commission’s release of a voluntary Code of Practice on Transparency of AI-Generated Content. Providers and deployers may choose to adhere to the code of practice as a means of facilitating and demonstrating compliance with the marking and labelling requirements under Article 50.
Alongside the transparency obligations imposed by the AI Act, developers and deployers of AI systems must ensure strict adherence to broader data protection rules. In particular, the General Data Protection Regulation (GDPR) regulates the processing of any personal data through an AI system.
Enforcement under the EU AI Act
2 August 2026 also marks the point at which the AI Act’s supervisory framework became operational. National market surveillance authorities are primarily responsible for the supervision and enforcement of most obligations under the AI Act, whilst the European Commission, assisted by the AI Office, is entrusted with the supervision and enforcement of obligations applicable to providers of general-purpose AI models under Chapter V of the AI Act.
As a result of the application of Article 50, failure to comply with the transparency obligations may be subject to administrative fines of up to €15,000,000 or 3% of the total worldwide annual turnover, whichever is higher. Moreover, providers of general-purpose AI models may likewise face fines of up to €15,000,000 or 3% of the total worldwide annual turnover, whichever is higher, for infringements of the obligations applicable under Chapter V of the AI Act.
In addition, penalties may be imposed on providers of general-purpose AI models for failure to comply with requests for information, failure to cooperate with supervisory measures, or failure to provide access to a general-purpose AI model for the purposes of evaluation and regulatory oversight.
Key Takeaways
As of 2 August 2026, transparency obligations under Article 50 of the AI Act have become applicable. Providers and deployers of AI systems should therefore ensure that they are in line with the new transparency obligations, as these are now subject to active regulatory oversight and enforcement.
Providers of general-purpose AI models should remain mindful of their existing obligations under Chapter V of the AI Act which became applicable on 2 Augst 2025, as well as the enhanced supervisory and enforcement powers available to the Commission and AI Office as of 2 August 2026.
Although the Digital Omnibus on AI postpones the application of various key requirements until December 2027, organisations should not consider this to be a relaxation of the regulatory framework. Rather, the extended implementation period presents an opportunity to undertake the necessary preparatory work to ensure readiness once those obligations become applicable.
