The AI Act: Simplified Rules, Ongoing Challenges

The AI Act creates the first legal framework for the use of AI worldwide. Shortly before important parts of the regulation entered into force, the European Union presented significant simplifying amendments it via the so-called AI Omnibus, offering some relief for companies in Bavaria.

Key elements of the AI Act

The AI Act follows a risk-based approach. It differentiates between AI system with unacceptable, high, limited and minimal risk. Systems posing unacceptable risks to health, safety and fundamental rights, are prohibited e.g. systems for emotion recognition in workplaces and for social scoring. AI systems classified as high-risk, have to be brought into compliance with the requirements of the AI Act. This concerns AI tools for employment and management of workers as well as systems used in critical infrastructures. Companies must fulfill requirements for risk management, data quality as well as documentation and human oversight.

Since August 2, 2026, some basic obligations apply to all providers and operators of AI systems, including transparency requirements. Anyone using customer chatbots or publishing AI-generated, unverified text, images, and videos must label them clearly for users. In addition, employers must take appropriate steps to promote the AI literacy of employees who work with AI systems. Importantly, the AI Omnibus now clarifies that companies are not required to guarantee any specific minimum level of proficiency for each individual worker.

Effects of the AI Omnibus

The AI Omnibus, adopted in July, streamlines some of the requirements, thereby simplifiying the AI Act. It extends implementation deadlines for high-risk AI and grants simplifications for small and medium-sized enterprises as well as to so-called small mid-caps. Moreover, the AI Omnibus delineates the AI Act and specific product rules more clearly e.g. regarding medical devices. Finally, the AI Omnibus clarifies the responsibilities of EU and national authorities when it comes to oversight and implementation. These amendments give companies more time to prepare for the implementation of the AI Act, they reduce bureaucratic burdens and facilitate the implementation in daily business operations.

Even with the AI Omnibus, the rules remain complex and challenging. The AI Act will be an obstacle to conquer for the widespread use of AI by companies. It is crucial that the announced guidelines and implementing acts are presented soon. They must ensure a coherent application with as little red tape as possible.