Article 50 in Practice (in German)

Description

Generative AI is part of everyday work, from marketing images to chatbots to automated publications. Since August 2, 2026, the transparency requirements of the AI Act have applied directly to these use cases; the Code of Practice on Transparency of AI-Generated Content (June 2026) specifies how they are to be implemented. The focus is on the operator obligations under Article 50, paragraphs 3 and 4: emotion recognition and biometric categorization must be disclosed, as must deepfakes and texts on matters of public interest, with certain exceptions.

This training translates the legal text into concrete decisions using realistic case studies: When must content be labeled? Who is responsible—the provider or the operator? Participants learn to classify their own AI use cases, apply appropriate labeling methods, and establish internal processes. They leave the training with a clear understanding of the transparency obligations and a practical path to compliance.

Learn objectives

• You can correctly classify typical AI use cases and determine whether—and, if so, which—operator obligations under Article 50(3) or (4) of the AI Act apply, including the existing exceptions.
• You can determine the division of responsibility between providers and operators and classify how provider obligations under Article 50(1) and (2) should be assessed when purchasing AI systems.
• You will plan concrete initial implementation steps: select appropriate labeling practices, establish internal processes to fulfill your organization’s operator obligations, and assess whether signing the Code of Practice would be beneficial for your organization.

Target audience

This training is designed for anyone who uses generative AI in their operations or is responsible for its implementation: founders, product and marketing leads, compliance and data protection officers, as well as communications and digital transformation teams in government agencies and local municipalities. It is equally suitable for startups, SMEs, and the public sector.

Prerequisites

No prior legal training is required. A general understanding of the AI Act, as well as personal examples from your daily work with generative AI—particularly as an operator—are beneficial but not required.

Trainer

Claudia Baumgartner, Trustworthy AI Expert, appliedAI Institute for Europe.

Many thanks to the community partners of the Bavarian AI Innovation Accelerator for their support in communicating and promoting the project offerings: BIHK – Bavarian Association of Chambers of Industry and Commerce, e.V., vbw – Bavarian Industry Association, e.V., Bayern Innovativ GmbH, UnternehmerTUM GmbH, Munich Innovation Ecosystem GmbH, fortiss GmbH, German Mittelstand e.V., and BAIOSPHERE.

This training is offered as part of the Bavarian AI Innovation Accelerator. The project is funded by the Bavarian State Ministry for Digital Affairs. Details: https://innovationsbeschleuniger.bayern/