The EU AI Act, formally known as the Artificial Intelligence Act, establishes a detailed regulatory framework to govern AI technologies across Europe. The Act classifies AI systems into categories based on risk—unacceptable, high, and low—imposing different requirements for each. Enforcement will start with a staggered timeline beginning in "November 2024 expanding through late 2027, with organizations potentially facing fines up to €35 million or 7%" of their global turnover for non-compliance.​ 

As noted in The Gartner® Report, 2024 Gartner® Getting Ready for the EU AI Act, Phase 1: Discover & Catalog, "almost every organization has exposure to the AI Act because you are responsible for not only the AI capabilities you build, but also those capabilities you already bought." This is crucial as a Gartner survey reveals that "As far back as 2021, more than 80% of organizations using AI deployed hundreds of AI models", many of which are embedded in SaaS platforms and may not be easily visible. Consequently, security, data, and risk management leaders must prepare for compliance by addressing both internally developed and externally sourced AI systems, ensuring they meet the Act’s rigorous standards to avoid substantial penalties and legal risks. Download the Gartner® Report to understand AI-enabled capabilities ahead of the mandatory risk assessment.​ 

Key findings include:​ 

• Risk categories and requirements defined in the EU AI Act​ 
• The projected timeline for the enforcement of the EU AI Act rules​ 
• AI-enabled use cases and the four classes of AI adoption​ 
• Methods for cataloging AI-enabled cases with sufficient detail​ 

Gartner, Inc. Getting Ready for the EU AI Act, Phase 1: Discover & Catalog. Nader Henein, Gabriele Rigon. 2 April 2024.