Optimal Sequencing of Projects with Uncertain Regulatory Costs; Ensuring Transparency and Reducing Bias in AI

This research contributes a stochastic optimization model that is operationalized for a specific application, through the uncertainty of random inputs by expertly assessed quantiles. These assessments are used to estimate parametric distributions from selected families, enabling complete quantification of uncertainty. Where available, historical data supplements the expert’s judgment—statistical analysis of carbon offset futures provides credible intervals to guide quantile assessment for the cost of mitigation instruments.

Packaged together with a mixed-integer linear program, the optimization model integrates these uncertainty inputs and selects an optimal sequence of projects and mitigation instruments to achieve the specified regulatory target. This allows CapTech, the company funding this research, to enable its clients with a decision-support tool to combine empirical data with expert assessment, providing organizations with a framework for project planning in a regulatory environment.

To ensure fairness and transparency without stifling innovation, regulatory bodies must evolve at a pace that keeps up with rapid technological advancements in AI. Establishing structured, adaptable guidelines is essential for promoting ethical AI development. The integration of RRI with current regulatory orders and academic literature allows for a comprehensive analysis of AI regulation, providing insights into how these frameworks can better account for the ethical, social, and technological challenges posed by AI.

It is crucial to adopt a holistic approach that involves improving data quality, increasing diversity in development teams and throughout the AI lifecycle, ensuring transparency in AI decision-making through explainable AI, and integrating robust regulatory frameworks. Government agencies like the FTC and EEOC, under the current legal landscape, will play pivotal roles in enforcing fairness in AI. Furthermore, the DOJ must develop a dedicated AI unit to address and monitor malpractice as the legal landscape evolves to be best prepared to handle this rapidly evolving technology. By drawing on frameworks like RRI, this research aims to contribute to the development of comprehensive, adaptable guidelines that promote fairness while fostering innovation.

School of Engineering and Applied Science

Bachelor of Science in Systems Engineering

Technical Advisor: Roman Krzysztofowicz

STS Advisor: Pedro Francisco

Technical Team Members: Christopher Woods, Mitch Mitchell