2026-06-05 ジョージア工科大学
＜関連情報＞
- https://research.gatech.edu/involving-communities-model-design-could-reduce-bias-ai
- https://link.springer.com/chapter/10.1007/978-3-032-15283-1_13
AIがシステムや組織全体で使用される際に、参加型モデリングがいかに集団的バイアスの軽減を可能にするか How Participatory Modeling Can Enable Collective Bias Mitigation when AI Is Used across Systems and Institutions
Erik W. Johnston & Reeham R. Mohammed
Participatory Modelling and Simulation to Improve AI-based Public Social Services Published:26 March 2026
DOI:https://doi.org/10.1007/978-3-032-15283-1_13
Abstract
The rapid adoption of AI in governmental contexts is outpacing concurrent research efforts to promote responsible innovation. In the context of public institutions, the application of AI technologies has been framed as a potential means to minimize systemic bias and enhance the effectiveness of public service delivery (Mikhaylov et al., 2018). Despite extensive equity-focused research, real-world deployments still fall short of best-practice models, widening the theory–practice gap (Yigitcanlar et al., 2024). In research triangulated between empirical research of participatory modeling of systems of care in Peoria, Illinois, analyses of 311 non-emergency systems in Boston, and stakeholder interviews in institutions of public higher education, participatory methods revealed the complex interdependencies of civic infrastructures, the pathways through which bias is introduced and propagated, and the promise of biomimetic design principles—particularly those inspired by immunological analogies—to inform anticipatory, adaptive, and preventative interventions.
