Duke’s Society-Centered AI (SCAI) initiative hosted its annual hackathon, bringing together students, faculty, and practitioners to develop innovative solutions that apply artificial intelligence in socially responsible ways. The event challenged participants to design projects that address real-world problems while prioritizing equity, transparency, and public impact.
Among the finalists was our team representing the Duke Master in Interdisciplinary Data Science (MIDS) program. The team (Eric Ortega Rodriguez, Ailina Aniwan, and Ilseop Lee) 
advanced to the final round, where they presented their proposal to judges Hazem Mahmoud, George Dohanich, and Funda Gunes, Ph.D.
Their project, EqualPaws, explored the growing use of photo scoring and recommendation systems in animal shelters. As shelters increasingly rely on AI-driven tools to evaluate pet photos and recommend animals to prospective adopters, the team examined how these systems may unintentionally disadvantage certain groups, particularly senior or disabled pets.
To address this challenge, the team developed a benchmarking framework designed to evaluate whether such systems promote equitable adoption outcomes. The framework includes metrics that assess whether different groups of animals receive fair access to high photo scores, proportional visibility in ranked recommendations, and comparable benefits from photo optimization tools. Using a counterfactual approach, the model measures whether system adjustments meaningfully improve visibility and opportunity across pet groups.
The hackathon, hosted by Brinnae Bent, Ph.D., Christopher Bail. Ph.D., and the SCAI team, emphasized innovation at the intersection of technology and society. MIDS student participation highlighted the program’s commitment to developing data scientists who are equipped not only with technical expertise, but also with a strong foundation in ethical and socially responsible AI.
The event showed the growing role of interdisciplinary data science in serving the public good.