ProductsModulesPricingBlogAbout PL Log in See demo
Module / AI Act

AI Act. AI system register with FRIA and human oversight.

Regulation 2024/1689 (AI Act) classifies AI systems into 4 risk levels. High-risk systems require a FRIA (Fundamental Rights Impact Assessment), human oversight, and transparency. RealSec handles all of this in one module without an Excel marathon.

01 / AI System Register

Risk classification: minimal / limited / high / unacceptable.

Every AI system in the organisation (built or procured) requires classification. Annexes I-III of the AI Act determine high-risk. Unacceptable = deployment blocked. Limited = transparency obligation. Minimal = best practice.

  • CLASS4 levels: minimal / limited / high / unacc.
  • ANNEXESAnnex I-III mapped to use cases
  • AI ACTArt. 6, 9, 10, 11 for high-risk
02 / FRIA for High-Risk

Fundamental Rights Impact Assessment with questionnaire.

Required for all high-risk systems deployed by public bodies and certain private ones. Questionnaire based on Art. 27 AI Act. The outcome determines additional obligations.

  • ART. 27Full FRIA questionnaire
  • CONSULTConsultation with Ombudsman / NGO (when required)
  • PUBLICPart of results published
03 / Human Oversight + AI Literacy

Art. 14 + Art. 4: oversight + user training.

Every high-risk system must have designated human-in-the-loop reviewers. AI Literacy training for all employees who interact with AI. Completion rate tracking per department.

  • OVERSIGHTDesignated reviewers with RACI
  • LITERACYTraining + completion tracking
  • INCIDENTWorkflow when AI produced an error
04 / Logging + Transparency

Art. 12 + Art. 13: lifecycle logs + transparency.

Automatic logging throughout the entire lifecycle of a high-risk system: inputs, decisions, human corrections. Transparency notice for end users: "this system uses AI, here is what it does".

  • LOGSAuto-logging compliant with Art. 12
  • NOTICETransparency templates per use case
FAQ

Najczęstsze pytania

Providers and deployers of AI systems in the EU, as well as companies that merely use AI in their operations (e.g. candidate scoring, model-based automation, chatbots). Obligations depend on the role and the risk class of the system.
Four: unacceptable (prohibited, e.g. social scoring), high (e.g. recruitment, credit scoring), limited (transparency obligation, e.g. chatbots), and minimal (no additional requirements).
For prohibited practices up to EUR 35 million or 7% of total annual worldwide turnover. Fines for breaching obligations for high-risk systems are lower but still in the millions of euros and as a percentage of turnover.
Fundamental Rights Impact Assessment (Art. 27 AI Act), required for high-risk systems deployed by public bodies and certain private entities.
Prohibitions on unacceptable practices have been enforceable since February 2025, rules for general-purpose AI models since August 2025, and most obligations for high-risk systems from 2 August 2026.

AI Act without the magic.

Demo on real data: system classification, FRIA workflow, AI Literacy tracker.

Umów demo →