Traceability and accountability by construction
Published in International Symposium on Leveraging Applications of Formal Methods, 2024
Recommended citation: Wenzel, J., Köhl, M. A., Sterz, S., Zhang, H., Schmidt, A., Fetzer, C., & Hermanns, H. (2024, October). Traceability and accountability by construction. In International Symposium on Leveraging Applications of Formal Methods (pp. 258-280). Cham: Springer Nature Switzerland. https://link.springer.com/chapter/10.1007/978-3-031-75387-9_16
As artificial intelligence (AI) systems influence ever more high-stake decision-making processes, such as university applicant screening or medical diagnoses, ensuring the trustworthiness of these systems and their decisions is crucial. This paper presents a significant step towards achieving trustworthy AI decisions by introducing a novel framework for enhancing traceability and accountability by construction. Our approach encompasses the entire decision-making pathway—from the raw datasets used to train the AI system, through the algorithms and programs employed, to the involved parties and the final decisions made. At the core of our methodology is the Decision Bill of Materials (DBOM), which meticulously documents all elements contributing to a decision while ensuring accountability and traceability through cryptographic signatures. Furthermore, we leverage results from logic programming to enable systematic reasoning about the processes and decision documented in a DBOM. This allows us to verify that the system meets specific certification standards and that individual decisions can be qualified as trustworthy. This framework not only advances the construction of reliable AI systems but also aligns technological developments with ethical imperatives and regulatory expectations.