Collaborative Design of Data and AI Systems for Science and Society
June 12-13, Central Campus Classroom Building — Ann Arbor, MI
The Research and Development Strategic Visioning (RDSV) team recently hosted its second workshop on Collaborative Design of Data and AI Systems for Science and Society with a diverse group of experts from U-M, Microsoft and Detroit.
Public-private research partnerships are an important component of the work at MIDAS. Such partnerships are all the more important at the moment when AI promises to fundamentally transform science and society and organizations across sectors are all exploring what this means to their organizations and the people in their community. Research collaborations, however, have mostly happened by chance, depending on random factors such as who one happens to meet at a conference. The RDSV project team aims to build a prototype for intentional and structured collaboration development across sectors. e.
This RDSV project served as a testbed for a new approach. The project team designed facilitation strategies to bring new collaborators together, share their vision, identify common interests, and design research projects that are meaningful for all involved. The facilitation process itself is a core part of the story: adaptive, reflective, and focused on building durable, productive partnerships. Through this work, the RDSV team is shaping a model for how public and private sectors can collaborate meaningfully and navigate complex, high-impact challenges.
The workshop participants built on the research project ideas that they developed in the first workshop and determined the scope, deliverables, datasets and technical expertise needed, and key collaborators for three projects:
- Chatbots to Enhance City Services
- Flood and Erosion Risk Policy Analysis Tool
- Computer Vision for City Planning
The remarkable progress excited all participants. Even though the RDSV team designed the first two workshops, the participants took it upon themselves to design the next workshop, which will consist of a hackathon to build prototypes for all three projects, and a presentation to key stakeholders to seek resources for full-scale project implementation. They also recommended that MIDAS use the structured approach developed through the RDSV project to continue building collaboration on responsible and ethical AI.
Dr. Jing Liu, the PI of the RDSV project, noted that the different but complementary senses of priority that participants brought to the table made the three projects feasible and exciting: “Detroit participants led the discussion to ensure that the intended research outcomes are meaningful. U-M researchers provided expertise to ensure the research rigor and feasibility. Microsoft engineers proposed an implementation plan that was astonishingly efficient. This is exactly why such collaboration can push research to the next level.”
