ETD

• Consumer protection
• Digital advertising
• Generative AI

Digital advertising law has traditionally relied on structural separability: sponsored content is presumed identifiable because it originates from an accountable author whose intent can be traced. Large language model (LLM) interfaces complicate this assumption, because commercial influence can shape which advertisements are selected alongside a response, or how advertising is personalised across a platform ecosystem based on conversational data.

This thesis examines whether the European Union's transparency and advertising disclosure rules remain adequate to confront this shift. For this, the thesis conducts a doctrinal analysis of the Unfair Commercial Practices Directive, the Digital Services Act, the AI Act, and the GDPR, assessing their capacity to reach these configurations and finding that disclosure obligations attach unevenly. From this, it develops a graduated governance framework, distinguishing ordinary disclosure cases from those raising user vulnerability or informational asymmetry, and identifies possible mechanisms suited to each.