EU AI Act

144 Pages condensed to One LawMap

Jurisdiction

European Union

Primary Rules

EU Regulation 2024/1689 (EU Artificial Intelligence Act)

Target Audience

All

Project Hash ID

#V065

Project Category Name

Athena

Regulation (EU) 2024/1689, the EU Artificial Intelligence Act, is the world's first comprehensive legal framework for AI. It applies to any provider, deployer, or importer of AI systems operating in the EU, regardless of where they are headquartered. The Act takes a risk-based approach: the stricter the potential harm an AI system can cause, the more rigorous the legal obligations it attracts. At the extreme, certain AI applications are outright prohibited. High-risk systems, such as those used in healthcare, law enforcement or employment, must pass conformity assessments, obtain CE marking and maintain comprehensive documentation. General-purpose AI models like large language models face transparency and systemic risk obligations scaled to their computational power. Lower-risk systems face lighter-touch transparency requirements or purely voluntary codes of conduct.

The LawMap condenses a highly technical, 144-page regulation into a single scannable map. For legal, compliance, and business teams, it immediately answers the two most practically relevant questions, does my AI system fall under the Act, and if so, what must I do, without requiring full reading of the legislative text.

Work Steps

Read and re-read the Act

Multiple readings of Regulation (EU) 2024/1689 were necessary to move from surface familiarity to structural understanding. Early readings focused on the overall architecture; later readings zeroed in on specific articles, annexes, and definitions. Summary notes were taken after each pass, gradually distilling the Act's logic into its core dimensions: risk tiers, classification criteria, and obligations.

Identify core organizing principle

After sufficient immersion in the text, the risk-based tiered structure emerged as the natural backbone of any visualization. A conscious decision was made to treat risk level as the primary axis, with criteria and obligations as secondary dimensions, reflecting the Act's own internal logic.

Select optimal visualization

Several types of visualizations were evaluated: decision trees, hierarchical diagrams and matrix tables. Given that the content has three parallel dimensions across six categories, a matrix table was selected as the most information-dense and scannable format.

Hand-drawn draft

A rough sketch was used to test the three-column matrix layout before committing to a digital format. This step helped resolve early layout questions, such as how to handle the provider/deployer split within the High Risk row, and whether GPAI deserved one or two rows.

Design, tweak and produce the final LawMap

The matrix was laid out in a professional design tool, formatted for DIN A3.