AI Act-Oriented Risk Assessment Framework for High-Risk Intelligent Systems
Keywords:
AI Act, risk assessment, high-risk AI systems, governance, system architecture, fairness, robustness, sustainability, socio-technical systems, regulatory complianceAbstract
The European Union's Artificial Intelligence Act (AI Act) establishes a tiered regulatory framework that imposes stringent requirements on high-risk AI systems, necessitating robust and transparent risk assessment methodologies. This paper presents a comprehensive system-level risk assessment framework specifically designed for high-risk intelligent systems as defined by the AI Act. The framework integrates technical, governance, and socio-technical dimensions to evaluate risks across the entire lifecycle of an AI system, from conception and training to deployment and continuous monitoring. We examine structural trade-offs between system performance and safety, architectural patterns that support compliance, and the infrastructure needed for ongoing validation. The treatment addresses issues of robustness, fairness, sustainability, and accountability, drawing on insights from large-scale systems engineering, regulatory science, and AI ethics. Through cross-domain comparisons and case illustrations, the framework is shown to be adaptable to sectors such as healthcare, critical infrastructure, criminal justice, and employment. The paper further explores policy implications and forward-looking perspectives on how such frameworks can evolve as AI capabilities and regulatory landscapes mature. This work aims to provide researchers, practitioners, and policymakers with a rigorous yet practical foundation for operationalizing AI Act requirements in the design and governance of high-risk intelligent systems.
References
1. European Commission. (2021). Proposal for a Regulation laying down harmonised rules on artificial intelligence (AI Act). COM(2021) 206 final.
2. IEEE. (2019). IEEE Standard for Software Quality Assurance Processes (IEEE Std 730-2014).
3. International Organization for Standardization. (2018). Risk management — Guidelines (ISO 31000:2018).
4. Crawley, E., Cameron, B., & Selva, D. (2015). System architecture: Strategy and product development for complex systems. Pearson.
5. Floridi, L., & Cowls, J. (2019). A unified framework of five principles for AI in society. Harvard Data Science Review, 1(1). https://doi.org/10.1162/99608f92.8cd550d1
6. European Parliament. (2024). Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence. Official Journal of the European Union.
7. Veale, M., & Zuiderveen Borgesius, F. (2021). Demystifying the draft EU Artificial Intelligence Act. Computer Law Review International, 22(4), 97-112.
8. Smuha, N. A. (2021). From a 'race to the bottom' to a 'race to the top'?: The European Union's approach to the regulation of artificial intelligence. European Journal of Risk Regulation, 12(1), 1-17.
9. European Commission. (2001). Directive 2001/95/EC on general product safety.
10. Ebers, M. (2021). Regulating AI in the EU: A new regulatory framework for artificial intelligence. Journal of Law and the Biosciences, 8(1), lsab005.
11. Rajpurkar, P., Chen, E., Banerjee, O., & Topol, E. J. (2022). AI in health and medicine. Nature Medicine, 28(1), 31-38.
12. Amodei, D., Olah, C., Steinhardt, J., Christiano, P., Schulman, J., & Mané, D. (2016). Concrete problems in AI safety. arXiv preprint arXiv:1606.06565.
13. Schiff, D., Biddle, J., Borenstein, J., & Laas, K. (2021). What's next for AI ethics, policy, and governance? A global overview. Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society, 153-158.
14. Hardt, M., Price, E., & Srebro, N. (2016). Equality of opportunity in supervised learning. Advances in Neural Information Processing Systems, 29.
15. Ananny, M., & Crawford, K. (2018). Seeing without knowing: Limitations of the transparency ideal and its application to algorithmic accountability. New Media & Society, 20(3), 973-989.
16. Green, B., & Chen, Y. (2019). The principles and limits of algorithm-in-the-loop decision making. Proceedings of the ACM on Human-Computer Interaction, 3(CSCW), 1-24.
17. Kaminski, M. E., & Rubinfeld, D. L. (2023). The regulation of AI in the European Union: A primer. Journal of European Competition Law & Practice, 14(2), 95-105.
18. Wachter, S., Mittelstadt, B., & Floridi, L. (2017). Why a right to explanation of automated decision-making does not exist in the General Data Protection Regulation. International Data Privacy Law, 7(2), 76-99.
19. Selbst, A. D., Boyd, D., Friedler, S. A., Venkatasubramanian, S., & Vertesi, J. (2019). Fairness and abstraction in sociotechnical systems. Proceedings of the Conference on Fairness, Accountability, and Transparency, 59-68.
20. Goodfellow, I. J., Shlens, J., & Szegedy, C. (2015). Explaining and harnessing adversarial examples. International Conference on Learning Representations.
21. Barocas, S., Hardt, M., & Narayanan, A. (2019). Fairness and machine learning: Limitations and opportunities. MIT Press.
22. Edwards, L., & Veale, M. (2017). Slave to the algorithm? Why a 'right to an explanation' is probably not the remedy you are looking for. Duke Law & Technology Review, 16(1), 18-84.
23. Bommasani, R., Hudson, D. A., Adeli, E., Altman, R., Arora, S., von Arx, S., ... & Liang, P. (2021). On the opportunities and risks of foundation models. arXiv preprint arXiv:2108.07258.
24. IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems. (2019). Ethically aligned design (2nd ed.). IEEE.
25. Jobin, A., Ienca, M., & Vayena, E. (2019). The global landscape of AI ethics guidelines. Nature Machine Intelligence, 1(9), 389-399.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2024 Computational Intelligence Systems
This work is licensed under a Creative Commons Attribution 4.0 International License.
This article is published under the Creative Commons Attribution 4.0 International License (CC BY 4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
