Now it’s final: The AI ​​law is coming (soon). The law officially comes into force on August 1st, but not all rules apply immediately.

The European Union has published the final text of the AI ​​law. This means that the last hurdle for the AI ​​law has been overcome. The law will officially come into force twenty days after its publication, on August 1st.

After a lengthy approval process that required several rounds, the European Council finally gave the green light on May 21. The text of the law had been in the works since 2021. The interpretation of the new legislation is based on categories, with AI systems being classified as “high” or “low” risk. Stricter rules apply to high-risk systems.

step by step

The AI ​​law will officially come into force on August 1, but the rules will come into force in several phases. An initial period of six months will ban the use of AI technology with “unacceptable” risk, such as systems that award “social scores” or the use of facial recognition software in public spaces.

Developers of “general” AI systems such as ChatGPT and Google Gemini should mark August 1, 2025 in their calendars. From then on, exactly twelve months after the AI ​​​​law comes into force, they must respect the obligation of transparency. This means that users should be shown a warning when they come into contact with AI.

The full legislation will not come into full effect until 36 months later, in 2027. The AI ​​Office must ensure that systems and their developers stay up to date. Violating these AI rules has serious consequences. Fines are a maximum of 35 million euros or 7 percent of global turnover. This depends on the type of violation and the size of the company.