The paper proposes seven principles for institutional design for autonomous AI
Start here
Petter Törnberg published an arXiv paper arguing that as autonomous agents interact in persistent groups, institutional design should become an independent focus of AI research. He proposes seven design principles, including reciprocal contribution, accountable authority, and human guidance, and proposes an experimental scheme for running agent groups under different institutional arrangements, introducing MoltJournal, a programmable, observable API-first journal.
Generated by AI from reports · Updated 2 days ago
Attention and developments
Attention, hourly
Now 29, -2% in 24 hours·Peak 32 (Nov 13 18:00)
Related research
This event's paper
- Designing Institutions for Autonomous AI
arXiv · 2026-11-12 · Deep reading
Progress timeline
- arXiv: 其他 AI 安全论文Designing Institutions for Autonomous AI
Petter Törnberg published a Perspective paper arguing that as autonomous agents interact in persistent groups, institutional design should become an independent focus of AI research. The paper combines institutional political economy, social epistemology, and organizational theory to propose seven design principles, covering reciprocal contribution, accountable authority, functional differentiation, epistemic pluralism, criticism and correction, human guidance, and institutional adaptation. The authors use machine science as an example to illustrate how autonomous research challenges institutions built around human capabilities and constraints, and propose an experimental research program in which matched groups of agents operate under different institutional arrangements; they also introduce MoltJournal, an API-first journal for autonomous research agents, as an example of programmable and observable institutions.
Related discussion
Attention trend
New sources per day
- Nov 13: 1 new sources (1 outlets, 0 accounts)
- Nov 14: 0 new sources (0 outlets, 0 accounts)
- Nov 15: 0 new sources (0 outlets, 0 accounts)