Specifying Computational AI Regulation Compliance: Blueprint for a New Research Domain

Bill Marino, Nicholas D. Lane

Keywords: AI regulation, AI safety

Abstract: This paper rests on the premise that AI systems and models will not be able to comply with AI regulations at the necessary speed and scale unless their compliance is enforced through algorithms that run across the life cycle of the AI, dynamically steering it towards compliance in the face of variable conditions. Despite their inevitability, the research community has yet to specify exactly how these “computational AI regulation compliance” algorithms should behave — or how we should measure their success. To fill this gap, we specify a set of design goals for such algorithms. In addition, we specify benchmarks that can be used to quantitatively measure how close they come to achieving those design goals. By delivering this blueprint, we hope to give shape to an important but uncrystallized new domain of research — and, in doing so, incite necessary investment in it

Submission Number: 46