A deterministic risk-scoring engine designed to operationalize the EDPB SPE Programme guidelines (2025) and the FRASP (Functional Rights AI Assessment & Scoring Protocol).
This tool provides a structured interface for evaluating the probability and severity of privacy and fundamental rights risks in AI systems. It transitions from qualitative descriptions to a quantitative scoring model while maintaining legal rigor.
- Probability Assessment: multi-factor scoring based on usage frequency, historical precedents, and system robustness.
- Severity Modeling: granular evaluation of impact on natural persons across 11 dimensions.
- Automated 'Stopper' Logic: integrated override that sets overall severity to Level 4 (Very Significant) if critical factors—such as human dignity or physical safety—are severely impacted.
- Risk Matrix Visualization: real-time mapping of current system status onto the probability/severity heatmap.
Developed to support DPOs and AI researchers in conducting robust impact assessments (DPIA/FRIA). The methodology strictly follows the March 2025 recommendations of the European Data Protection Board (EDPB) Pool of Experts.
- Logic: Rule-based scoring engine.
- Stack: Python, Streamlit, Plotly (Data Visualization).
The content of this repository are licensed under the CC BY-NC-SA 4.0 license.