SERC awards third round of seed grants to MIT researchers

The MIT Schwarzman College of Computing’s Social and Ethical Responsibilities of Computing (SERC) has awarded a third round of seed grants. These grants support MIT researchers in contributing to the responsible development and deployment of technology.

This year’s selected proposals covers a wide breath of topics, from equity in music to data center sustainability to AI governance. Eight projects, led by MIT researchers from eight departments across all five schools and the college, have been selected to receive up to $92,000 in funding.

“The sustained response of over 70 proposals in our third seed grant round highlights how deeply the MIT community cares about the social and ethical impacts of computing,” says Nikos Trichakis, associate dean of SERC and J.C. Penney Professor of Management. “With submissions spanning the entire Institute and evaluated by a broad faculty panel, the chosen projects highlight the research that MIT collectively believes will make the biggest difference in this critical space.”

The eight projects and research leads are:

• “Building Evidence Before Trust: Risk Discovery for AI Decision Support in High-Stakes Social Systems,” led by Chuchu Fan, proposes the development of budget-aware Bayesian learning framework for auditing AI decisions-support systems prior to deployment.
• “Agent-Assisted Model Specification for Democratic AI Governance,” led by Dylan Hadfield-Menell, proposes the development and evaluation of an AI-assisted tool for collaborative model specification authoring in governance applications.
• “Representational Equity in Music AI: Idiom-Attuned Corpora for Cross-Modal Learning,” led by Anna Huang, proposes a representational infrastructure and framework for responsible AI in cultural domains.
• “Democratic Governance in AI Infrastructure,” led by Jason Jackson, proposes a framework toward the governance of AI’s physical infrastructure and the communities that bear its costs.
• “Responsible AI Implementation in Healthcare Through Frontline Worker Voice,” led by Erin Kelly, proposes the development of a repeatable, worker-centered method for incorporating frontline workers’ perspectives in key AI implementations and decisions.
• “Organizational Sociology of AI: Understanding How and Why Multi-Agent Systems Fail,” led by Georg Rilinger, proposes an open-source research platform that implements key constructs from organizational theory as manipulable experimental variables.
• “Can A Market for AI-Crawling Keep the Open Web Alive?” led by Tobias Salz, proposes the development of a structural model to examine how generative AI changes the incentives of content production.
• “Integrated Sustainability Impacts of Data Centers,” led by Noelle Selin, proposes the advancement of methods to analyze the social and ecological impacts of AI data center development.