AI Act Evaluation Benchmark: An Open, Transparent, and Reproducible Evaluation Dataset for NLP and RAG Systems

TL;DR

This paper introduces an open, transparent, and reproducible dataset for evaluating NLP and RAG systems' compliance with the EU AI Act, enabling more efficient and accurate assessments of AI risk levels and obligations.

Contribution

The authors developed a novel dataset and methodology for evaluating NLP models against EU AI Act compliance tasks, utilizing large language models for grounded scenario generation.

Findings

Achieved 0.87 F1-score for prohibited scenarios

Achieved 0.85 F1-score for high-risk scenarios

Demonstrated effective use of language models for grounded data generation

Abstract

The rapid rollout of AI in heterogeneous public and societal sectors has subsequently escalated the need for compliance with regulatory standards and frameworks. The EU AI Act has emerged as a landmark in the regulatory landscape. The development of solutions that elicit the level of AI systems' compliance with such standards is often limited by the lack of resources, hindering the semi-automated or automated evaluation of their performance. This generates the need for manual work, which is often error-prone, resource-limited or limited to cases not clearly described by the regulation. This paper presents an open, transparent, and reproducible method of creating a resource that facilitates the evaluation of NLP models with a strong focus on RAG systems. We have developed a dataset that contain the tasks of risk-level classification, article retrieval, obligation generation, and question-answering for the EU AI Act. The dataset files are in a machine-to-machine appropriate format. To generate the files, we utilise domain knowledge as an exegetical basis, combining with the processing and reasoning power of large language models to generate scenarios along with the respective tasks. Our methodology demonstrates a way to harness language models for grounded generation with high document relevancy. Besides, we overcome limitations such as navigating the decision boundaries of risk-levels that are not explicitly defined within the EU AI Act, such as limited and minimal cases. Finally, we demonstrate our dataset's effectiveness by evaluating a RAG-based solution that reaches 0.87 and 0.85 F1-score for prohibited and high-risk scenarios.