EU AI ACT
New obligations and what comes next
With the Digital Omnibus, the European Union has reshuffled the timeline and pushed back some high-risk obligations of the EU AI Act. However, August 2, 2026, remains an important date. For compliance officers, it’s worth taking a close look at what becomes binding now, and what has gained more time.
The AI Act as of August 2, 2026
As described in our post on the Code of Practice, deepfakes and AI-generated content on topics of public interest must be clearly labeled. Users must also be able to recognize when they are interacting with an AI system such as a chatbot. The goal is to prevent deception and strengthen public trust in sources and informational infrastructures.
The framework surrounding fines also takes effect as of this date. For the most serious violations, such as prohibited AI practices, fines of up to €35 million or 7% of global annual revenue are possible; for the remaining obligations, fines of up to €15 million or 3% could apply.
The Digital Omnibus Has Reorganized the Deadlines
Considering content, the requirements remain the same. Risk management, technical documentation, data governance, and human oversight are unchanged, only the date on which they will be enforced has shifted. The labeling requirement is also only partially delayed: machine-readable marking of AI-generated content (watermarking) applies to generative systems that came to market before August 2, 2026, only starting December 2, 2026. For systems introduced after that date, it applies immediately.
What is yet to come for Companies
Since February 2025, the AI literacy obligation and the ban on certain AI practices have been in effect; since August 2025, the requirements for general-purpose AI (GPAI) models have applied. On August 2, 2026, the transparency obligations and new rules for fines follow, and by the end of 2026, machine-readable labeling for existing generative systems. The high-risk obligations then take effect starting December 2, 2027, for high-risk systems in areas such as credit assessment or law enforcement, and starting August 2, 2028, for product-integrated systems such as medical devices or vehicles.
Who Implements the EU AI Act in Germany
In Germany, the Bundesnetzagentur (Federal Network Agency) is building out the oversight and support structure. It is setting up the central contact point under Article 70, preparing the notification process for high-risk systems and market surveillance, and planning a center for competence and coordination. Through its AI Service Desk, it offers concrete assistance, including the AI Compliance Compass, an online tool for assessing whether a system falls under the AI Act and which risk class it belongs to.
Our AI Compliance Solution helps you manage your AI systems in a central catalog, classify them according to the requirements of the EU AI Act, and document all evidence in an audit-proof manner. This way, you can reliably stay on top of current and upcoming deadlines.
Source: Bundesnetzagentur