MIDAS Future Leaders Summit highlights Responsible Data Science and AI
April 29, 2022
Data science and AI are having a significant impact on society in uncountable ways, leading to huge benefits in many cases. Yet, increasingly complex analytical pipelines working with poorly understood heterogeneous data sets can also give rise to harms in many ways, including deleterious systemic effects such as the magnification of disinformation or surveillance capitalism. There has been tremendous recent interest in understanding and managing these concerns. The 2022 Future Leaders Summit, a Michigan Institute for Data Science (MIDAS) event at the University of Michigan, brought together 30 outstanding Ph.D. and Postdoctoral trainees from universities across the United States, with the theme of “Responsible data science and AI”.
Since its inception in 2019, this event, in spite of COVID disruptions, has become a much-discussed highlight of the academic data science community. More than half of the attendees this year were women and / or underrepresented minorities, just like in previous years, which is also a hallmark of this event.
The two-day event featured vision talks from faculty mentors, research presentations by the trainees and a public-private partnership discussion. The presenters highlighted current research and explored future research directions in equity and fairness, particularly in automated decision making, the explainability of analytical results, reproducibility and replication of scientific results, and systemic issues; particularly those impacting marginalized populations. The research presentation topics were as varied as the trainees themselves ranging from developing new methods to detect use of Deepfake technology to predicting severity of climate change effects to investigating the illicit trade of cultural artifacts.
The event also offered the next generation of data science and AI research leaders opportunities to network with peers, and receive career mentoring from both faculty mentors and representatives from Microsoft and Rocket Companies.
A mini-symposium for Responsible Data Science and AI, open to the public, was also part of the Summit. Prominent research leaders, who also served as mentors of Summit attendees, gave vision talks:
- H.V. Jagadish (Director, Michigan Institute for Data Science): “Why Data Scientists Should Care About Equity”
- David Mongeau (Founding Director & Professor of Practice, School of Data Science, The University of Texas San Antonio): “Responsible data science is equitable, informed, and secure”
- Shashi Shekar (Professor of Computer Science, University of Minnesota): “GeoAI and Spatial Data Science: with Great Power comes Great Responsibility”
- Michael Wellman (Professor and Chair of Computer Science and Engineering, University of Michigan): “When Algorithms Trade: Modeling AI in Financial Markets”
- Frauke Kreuter (Professor of Survey Methodology, Director of Social Science Data Center, University of Maryland): “Who Decides What Counts? AI and Big Data: Applications in Economic and Social Science Research”
“Big data is all around us and it’s important that we are intentional about how it is used and aware of what could be missing,” said Kreuter, “An event like this Summit is important not just because I can share my experiences and mentor these up and coming researchers, but also because senior faculty like myself can learn so much from them as well.”
The Summit was capped off with presentations from Ethical AI researchers at Rocket Companies and Microsoft, and a roundtable discussion on public-private partnership to advance the ethical use of data science.
“Companies like Microsoft have huge responsibilities to understand the risks of the systems they are developing but also learning to control those risks from a real-world perspective ,” said Microsoft’s Ece Kamar, who is at Microsoft Research and a technical advisor of Microsoft’s committee on AI, Ethics and Effects in Engineering and Research (AETHER).
“Our approach to responsible and ethical data science and AI starts with a motto of ‘Do the right thing.’ We were inspired to lean into these efforts from a presentation that Dr. Jagadish gave about equitable data science and we are excited to share what we’ve accomplished in the last 16 months with this outstanding group,” said Brian Stucky, Lead for Rocket Ethical AI at Rocket Companies.
“I enjoyed the diverse voices from the faculty mentors, industry speakers, and my fellow Summit participants’ perspectives.”said Yuhan ‘Douglas’ Rao, a postdoctoral researcher at the North Carolina Institute for Climate Studies, “The discussions were very engaging; it’s rare to have such broad exposure to different disciplines in regular academic training .”
“These trainees are the next generation of academic leaders in data science. We wanted to acknowledge their efforts and nurture them to think ethically and responsibly about data science as they begin their careers in an ever changing and quickly expanding discipline,” said Jagadish. “I speak for the mentors and industry leaders attending the summit when I say the future of data science is in capable hands.”