Abstract: Artificial intelligence is reshaping how we learn, work, and live, forcing universities to confront an existential challenge. If machines can perform the tasks of knowledge workers more cheaply and efficiently, what is the value of a university education — and who will continue to pay for it? At the same time, as expertise itself is questioned and public trust in institutions erodes, can universities sustain their legitimacy as bastions of knowledge and civic authority? This talk will explore these twin questions of the enduring (?) value and legitimacy of universities.

Abstract: Advances in self-supervised AI revolutionized modern machine intelligence, but important challenges remain when applying these solutions in IoT contexts - specifically, on lower-end distributed embedded devices with multimodal specialized sensors, where ample training data are not readily available. The talk discusses challenges in offering self-supervised machine intelligence services to support distributed embedded sensing applications. The intersection of IoT applications, real-time requirements, distribution challenges, and self-supervised AI motivates several important research directions. For example, how to adapt self-supervised training pipelines to the embedded sensing domain? Can one develop foundation models for IoT that offer extended inference capabilities from multimodal time-series data? How to endow these models with an understanding of space (and spatial signal propagation) in order for them to reconstruct the state of the physical environment from multiple distributed sensor observations? How to overcome the challenge of data scarcity when it comes to training such AI models with specialized sensor data that are not as widely available as text and images? The paper addresses the above questions and presents initial empirical results on using the answers to train small foundation models for embedded sensor data.

Abstract: Current AI is designed as a rough emulation of human intelligence.  If instead we designed AI to complement human intelligence we can achieve much more useful performance.  I will show examples from finance, science, health, patents, and policy.

Abstract: In the first part of this talk, I will briefly discuss how technologies for communication and computation—including most recently AI—are now making it possible for us to create new kinds of superminds that include both people and computers doing things that were never possible before. This part of the talk will include some of the background work and underlying philosophy that led to the M3S T6 project on “Designing human-AI teams.” The last part of the talk will be a chance for the audience to ask me anything they want to about the lessons I have learned in a career as a professor.

Description: Andrés shares his insights and experience on the Institutional Review Board (IRB) process and working with human subjects. This informative session is designed to help researchers navigate experimental studies involving human participants. Andrés explains why special ethical guidelines are necessary for human research and provides a brief history of these guidelines. He then discusses the Belmont Report, the IRB process, and shares practical tips for IRB applications.

Abstract: As AI becomes more advanced, there is talk of AI agents: essentially, bots that execute tasks, both small and large. This should be no surprise given the several decade effort to make automobiles autonomous. I will discuss agency philosophically, and present a framework for thinking about agents and agency. I will also talk about errors in AI and human-AI co-work. At the same time, I will talk about human education, and the role of agency in learning at various levels of human development. I will contrast our efforts and investments in human and machine learning, and explore how we might reconcile the two.