The entry into force of Article 50 of the AI Act marks a key milestone in the process of regulating artificial intelligence in Europe. The regulation introduces a structured framework for transparency, establishing precise standards to make content generated or manipulated by synthetic systems recognizable and to clarify interactions between users and algorithms.

For companies that use or integrate AI-based solutions, understanding the scope of Article 50 allows them to accurately identify the necessary checks, distinguishing actual operational requirements from aspects that do not fall under this specific regulation.

What is Section 50 of the AI Act, and to whom does it apply?

Article 50 of the AI Act is the section of the European regulation dedicated to information and transparency requirements regarding the use of artificial intelligence.

The regulation identifies two distinct roles within the technology supply chain:

  • Providers: Organizations that develop, customize, or market AI systems in the European market.
  • Deployers: companies that use AI systems as part of their professional activities or operational processes.
What is AI ACT?

Providers and Deployers: The Division of Responsibilities

The policy assigns responsibilities depending on the organization’s role:

  • As a provider, the organization must design its systems so that interactions with AI are transparent (if not already evident from the context) and ensure that the outputs (text, images, audio, video) include a technically detectable watermark.
  • As a Deployer, the company is required to report synthetic content (deepfakes), specify when text published for informational purposes is generated by AI, and provide adequate information to users when using emotion recognition or biometric categorization systems.

Timing of the Notice: The deployer’s obligation to provide information must be fulfilled no later than the time of the user’s or recipient’s first interaction with or exposure to the AI system.

5-Point Operational Checklist for Businesses

Compliance with Article 50 requires an analysis of specific touchpoints where AI interacts with people or generates content.

The verification process can be broken down into five steps:

  1. Mapping Direct Interactions: Identify use cases in which AI interacts directly with individuals or generates content intended for external audiences, clearly distinguishing between the roles of provider and deployer.
  2. Virtual Assistants and Chatbots: Ensure that conversational interfaces or support services explicitly inform users that they are interacting with an automated system.
  3. Communication and External Channels: Provide a separate disclosure statement for images, audio, video, or text created using generative AI tools and intended for public consumption.
  4. Relations with AI software providers: Request formal confirmation from technology solution providers regarding the compliance of their modules with the technical marking requirements set forth in Article 50.
  5. Biometric or behavioral systems: If such systems are in use within the organization (for example, at access points or in security), verify that the relevant prior notice is in place and that the systems comply with the applicable regulatory framework (including the GDPR).

Official Documents of the European Commission

To facilitate technical understanding and the application of the standards, The European authorities have released two official documents:

  • The Code of Practice on Transparency of AI-Generated Content¹: a code of conduct published in June 2026 that outlines best practices for managing labeling and transparency.
  • The European Commission’s Guidelines on Article 50²: official guidelines issued in July 2026 that clarify the scope, exceptions, and examples of application for providers and deployers.

Clarification on Timelines: The Scope of the Digital Omnibus

With regard to the implementation schedule, it is important to clarify the scope of the postponement to December 2, 2026, introduced by the Digital Omnibus.

The measure has a limited and well-defined scope:

  • applies exclusively tothe technical requirement for output marking;
  • This applies only to providers of generative systems that were already on the market before August 2, 2026.

All other obligations under Article 50 applicable to deployers—including notices to users of virtual assistants and the labeling of synthetic content—will take effect on August 2, 2026.

The Integration of AI into Business Processes

The introduction to Article 50 sets out clear guidelines for the adoption of artificial intelligence in organizations. Understanding the regulatory framework enables companies to move forward with their digital transformation initiatives with greater operational certainty, while integrating transparency requirements into standard business practices.

At Regesta, we support companies in integrating Artificial Intelligence into their business processes, developing technology projects designed to maximize operational efficiency and ensure full compliance with industry standards.

Would you like to learn how to implement AI in your company effectively and securely?

Contact us and tell us what you need

¹ Source: Code of Practice on Transparency of AI-Generated Content
² Source: European Commission Guidelines on Article 50