Rakshit Trivedi is a postdoctoral associate in the Algorithmic Alignment Group at the MIT Computer Science and Artificial Intelligence Laboratory (CSAIL), where he works with Dylan Hadfield-Menell. His research is on cooperative AI, AI alignment, and multi-agent safety, with a focus on building AI systems that can navigate the multi-agent ecosystem of human society and engender cooperation between humans, AI agents, and institutions (Source: https://rstrivedi.github.io/).
Background
Trivedi completed his PhD at the Georgia Institute of Technology, advised by Hongyuan Zha. He collaborates with David Parkes (Harvard), Gillian K. Hadfield (Johns Hopkins), and Joel Leibo and the multi-agent team at Google DeepMind. He was selected as a Kavli Fellow by the US National Academy of Sciences in 2024 and has co-organized the NeurIPS Melting Pot and Concordia contests on cooperative intelligence for language-model agents (Source: https://rstrivedi.github.io/).
Research and the Democratic Matrix essay
Trivedi is a co-author, with Gillian K. Hadfield and Dylan Hadfield-Menell, of Building AI for the Democratic Matrix (Knight Columbia, March 2026); see The Democratic Matrix, Normative Competence, and Building AI for the Democratic Matrix: A Technical Research Agenda for Normative Competence and Normative Institutions (Hadfield + Trivedi + Hadfield-Menell, Knight Columbia, March 3 2026).
His contribution to that work is the technical and multi-agent grounding. Hadfield supplies the legal-institutional theory (the Hadfield-Weingast account of normative social orders) and Hadfield-Menell the alignment framing, while Trivedi's research program blends multi-agent reinforcement learning, generative agents, and game-theoretic modeling of cooperation. According to the essay, this is what makes its central claim tractable as an engineering agenda: that an agent can be built to detect social sanctions, attribute them to behaviors, and adjust (the Normative Competence primitive), rather than merely having static values encoded. His earlier work on normative infrastructure and normative reasoning in AI agents, presented at the Knight Symposium on Artificial Intelligence and Democratic Freedoms, preceded the Democratic Matrix essay.
Relationships
- affiliated-with: MIT CSAIL (Algorithmic Alignment Group)
- co-author-with: Gillian K. Hadfield, Dylan Hadfield-Menell
- author-of: Building AI for the Democratic Matrix: A Technical Research Agenda for Normative Competence and Normative Institutions (Hadfield + Trivedi + Hadfield-Menell, Knight Columbia, March 3 2026)
- supports: The Democratic Matrix, Normative Competence, Classification Institutions
- related: Agentic AI, AI Alignment
Sources
- Building AI for the Democratic Matrix: A Technical Research Agenda for Normative Competence and Normative Institutions (Hadfield + Trivedi + Hadfield-Menell, Knight Columbia, March 3 2026) — Hadfield, Trivedi & Hadfield-Menell, Building AI for the Democratic Matrix, Knight Columbia, March 2026.
- (Source: https://rstrivedi.github.io/) — Trivedi's academic homepage (affiliation, research focus, biography).
Page created 2026-05-25 by the gap-identifier (gap type 3 — co-author of a tracked foundational source, referenced from The Democratic Matrix, Gillian K. Hadfield, and Dylan Hadfield-Menell with no entity page). Confidence medium.