The MIT Stephen A. Schwarzman College of Computing has granted seed funds to seven projects exploring the synergy between artificial intelligence and human-computer interaction to enhance modern workspaces, aiming for improved management and heightened productivity.
Backed by Andrew W. Houston ’05 and Dropbox Inc., these initiatives are interdisciplinary, uniting researchers from computing, social sciences, and management.
The seed grants serve as a catalyst for research that can pave the way for substantial advancements in this rapidly evolving realm, while also fostering a community around questions related to AI-augmented management.
The selected projects and their research leaders encompass:
LLMex: Realizing Vannevar Bush’s Memex Vision with Large Language Models
Led by Patti Maes from the Media Lab and David Karger from the Department of Electrical Engineering and Computer Science (EECS) and the Computer Science and Artificial Intelligence Laboratory (CSAIL), this project takes inspiration from Vannevar Bush’s Memex and aims to create, implement, and test memory prosthetics using large language models (LLMs).
The AI-driven system is intended to intelligently assist individuals in managing copious amounts of information, accelerating productivity, and reducing errors by autonomously documenting work actions and meetings. It will facilitate retrieval based on metadata and vague descriptions, as well as proactively suggest personalized, pertinent information based on the user’s current focus and context.
Utilizing AI Agents for Simulating Social Scenarios
Led by John Horton from the MIT Sloan School of Management and Jacob Andreas from EECS and CSAIL, this project envisions the simulation of policies, organizational structures, and communication tools with AI agents before actual implementation. Integrating modern LLM capabilities as computational models of humans lends greater realism and potentially enhanced predictability to the concept of social simulation.
Human Expertise in the Age of AI: Finding Balance
Manish Raghavan from MIT Sloan and EECS, and Devavrat Shah from EECS and the Laboratory for Information and Decision Systems, spearhead this project. It acknowledges the rise of machine learning, AI, and algorithmic decision aids, considering how algorithms might complement human decision-making across diverse scenarios. The project envisions a harmonious future where AI and algorithmic decision aids enhance rather than replace human expertise.
Embedding Generative AI in U.S. Healthcare Settings
Julie Shah from the Department of Aeronautics and Astronautics and CSAIL, Retsef Levi from MIT Sloan and the Operations Research Center, Kate Kellog from MIT Sloan, and Ben Armstrong from the Industrial Performance Center lead this initiative.
It seeks to address the burnout experienced by healthcare professionals due to increased administrative tasks associated with electronic health records and technology. The project aims to create a comprehensive framework to explore how generative AI technologies can boost productivity and improve job satisfaction in healthcare settings.
Democratizing Programming with Generative AI Augmented Software Tools
Led by Harold Abelson from EECS and CSAIL, Cynthia Breazeal from the Media Lab, and Eric Klopfer from the Comparative Media Studies/Writing, this project capitalizes on the advancements in generative AI. It aims to transform computing education for individuals with no prior technical training by creating software tools that significantly reduce the need for coding when developing applications, thus reshaping traditional career assumptions in the software domain.
Acquiring Expertise and Societal Productivity in the Era of AI
Spearheaded by David Atkin and Martin Beraja from the Department of Economics, along with Danielle Li from MIT Sloan, this project delves into how generative AI can enhance cognitive task performance. It explores the potential impact of AI on skill acquisition and productivity and seeks to identify policy interventions that can maximize societal gains from these technologies.
AI-Enhanced Onboarding and Support
Guided by Tim Kraska from EECS and CSAIL, and Christoph Paus from the Department of Physics, this project addresses the learning curve associated with accessing resources like LLMs. By developing LLM-powered onboarding and support systems, the project aims to enhance user experiences and streamline support team operations for efficient utilization of these resources._
(Source: MIT News)