Seven MIT academics win Schmidt Sciences AI2050 fellowship

Summary

Five faculty members and two alumni from MIT have been selected for the 2026 Schmidt Sciences AI2050 Scholars cohort. Each is leading a distinct research project centered on how artificial intelligence can be 'largely beneficial' to society by 2050.

Why it matters

AI2050 directs the private philanthropy of Eric and Wendy Schmidt toward long-term AI questions that traditional grant cycles often fail to adequately fund. The project portfolio spans optimization theory, democratic sampling, and enzyme discovery. This diversity signals a deliberate effort to link AI research to concrete societal outcomes. The fact that all five faculty scholars are affiliated with MIT's Laboratory for Information and Decision Systems also illustrates how a single institutional hub can shape the direction of boundary-pushing AI research.

Highlights

• The program's core premise—that AI will be 'largely beneficial'—is treated as an established fact rather than a hypothesis to be tested.
• The projects cover a wide range: effective model training, structured public input for legislative bodies, enzyme function prediction, mitigating the harms of AI dependency, and translating AI patterns into testable theories.
• All five faculty scholars are affiliated with MIT's Laboratory for Information and Decision Systems. This suggests the existence of a tightly clustered institutional pipeline for AI2050 funding.

Critical take

The text reads as a flat press release that omits funding amounts, selection criteria, or how these efforts differ from existing grant-supported work. It does not question the program's optimistic framing, presenting the 2050 utopia as a design constraint rather than a debatable assumption.