From Sunday, 2 August 2026, the transparency obligations under Article 50 of the EU AI Act will apply across the European Union. Companies of all sizes and in all sectors will, among other things, be required to inform users when they are interacting with a chatbot or voice assistant. New labelling and information obligations will also apply to AI-generated and AI-manipulated content.
“Transparency builds trust and is a fundamental prerequisite for the responsible use of Artificial Intelligence. However, binding obligations also require authoritative guidance and clear definitions before they take effect. The fact that the European Commission published its final guidelines just 13 days before the obligations became applicable is a foreseeable false start,” says Alexander Rabe, Managing Director of eco – Association of the Internet Industry.
Despite the guidelines, key questions remain unanswered. Robust and interoperable technical standards for labelling AI-generated content are still lacking. There is particular uncertainty in several grey areas, such as content that has been manipulated using AI but is not entirely AI-generated, as well as the precise form that notices in chatbots and AI assistants should take.
“Companies must not be assessed against technical standards that have yet to be defined consistently. During the initial phase, guidance, a harmonised interpretation across Europe and proportionality must take precedence over sanctions-driven enforcement. Otherwise, start-ups and small and medium-sized enterprises (SMEs) in particular will pay the price for shortcomings in the European legislative process,” says Rabe. “If Europe is to have a chance in the AI race, we need competitive rules for Artificial Intelligence.”
With regard to enforcement in Germany, eco welcomes the central role assigned to the German Federal Network Agency (BNetzA). However, it is essential that the requirements be applied consistently and in a manner that works in practice.
“Enforcement will determine whether the AI Act strengthens trust in AI or becomes an obstacle to investment and innovation. The national implementation of the AI Act must result neither in a maze of German authorities nor in a European regulatory patchwork,” says Rabe.
At the same time, eco warns that transparency obligations must not create a false sense of security when it comes to identifying content manipulated or altered using AI. “Transparency obligations are no substitute for the need to establish broad AI literacy among the general public,” says Alexander Rabe.
eco therefore calls for clear, technically feasible standards to be introduced as quickly as possible, together with a coordinated, legally certain and innovation-friendly approach to enforcement across Europe.
