How should higher-ed institutions prepare students for an automated future?

Suresh Venkatasubramanian,
professor in the School of Computing

Artificial intelligence and machine learning are changing the way we make decisions. But they’re also doing more— changing the very DNA of our society, affecting our laws, our politics, our interactions, and even how we know what we know.  What our students need more than ever is a new kind of education that blends humanistic and mechanistic thinking: that helps us understand what it means to be human and live in society in an algorithmically mediated world.

Matthew Sigman,
Peter J. Christine S. Stang Presidential Endowed Chair of Chemistry

Everyone should have a basic understanding of computer science and statistics. We have to start thinking about a core curriculum, such as a junior level course on, say, data science and chemistry with general features of the types of algorithms and statistical questions we should be asking.  I don’t understand all of the mathematics behind it. I don’t think you necessarily have to. You have to understand, though, what the algorithms actually are doing.

Arul Mishra,
David Eccles Professor of Marketing

Higher-ed institutions need to stop treating AI within the confines of specific departments or schools and start thinking of it as a second language—a language that is now being used to answer questions in areas as diverse as literature and astronomy.  If AI becomes a second language, it will not be taught separately but integrated into existing curriculum. This approach will prepare students for a future where more and more tasks across all professions will be performed by AI algorithms.

Dan Reed,
senior VP for academic affairs and former corporate VP at Microsoft

Every student, regardless of background, needs to understand how to extract basic insights from sets of data and how to appreciate and assess the validity of common statistical arguments. Put another way, data literacy is now an essential skill, a peer with quantitative reasoning and communication skills.  These skills are necessary to appreciate whether learned models and AI yield ethical, valid, and trustworthy predictions and behaviors.