In recent years, the European Union has established the world’s first comprehensive and unified legal framework for AI-based systems (Regulation (EU) 2024/1689). This is not merely a set of recommendations, but a binding legal act that affects both Member States and industry stakeholders. The regulation will become applicable in stages, with full application expected from August 2026.

Hungary is among the first countries to implement the provisions set out in the regulation, which poses a significant risk for unprepared market participants.

What does the EU AI Regulation regulate?

It classifies AI systems into the following risk categories, to which different legal requirements are assigned:

• Unacceptable risk, prohibited AI practices (from February 2025)
• High-risk AI systems (from August 2026)
• General-purpose systems, transparency risk (from August 2026)
• Minimal risk

What obligations do industry stakeholders have to face?

• Compliance: ensuring that AI-based systems meet the requirements applicable to their respective risk category
• Transparency: providing detailed documentation to authorities on the operation of AI-based products, as well as on the traceability of risks
• Risk management: the obligation to carry out legal analyses prior to deployment and during operation, and to prepare audit reports

The AI Act Evaluator developed by AITIA is an auditing tool that automatically performs risk assessment and compliance evaluation under the EU AI Act based on the uploaded technical and other documentation. By following the logical structure of the regulation, the system determines which AI Act risk category the given AI solution falls into, whether it involves prohibited practices, whether it qualifies as a high-risk AI system, and which role and obligations apply to the organization concerned. The result is a transparent report that clearly presents the risk classification and the relevant conclusions.

With the help of the tool, developers as well as compliance and legal teams can quickly and consistently assess the compliance of their AI systems, significantly reducing the time and resource demands of manual analysis. The solution is particularly useful for organizations that manage multiple AI systems or wish to align complex documentation with the requirements of the AI Act in an automated and reliable manner.

AI Act Assessment

The system uses a predefined decision tree based on the AI Act, which:

• Provides consistent and compliance-relevant findings.
• Attaches a justification to each assessment point.
• Indicates the level of certainty.
• Identifies gaps in the documentation.

Areas of Application

• Preparing AI products and services for market launch.
• Supporting internal or external compliance audits.
• Accelerating the work of lawyers, compliance professionals, and development teams.
• Reviewing documentation of prohibited and high-risk AI systems.
• Preliminary risk analysis of suppliers’ and partners’ AI solutions.

Audit Reports

During the analysis process, the system generates documents that transparently capture the entire evaluation process.