Sources

Survey the field.

We have catalogued a variety of papers on LAS through community contributions and our work at Gigascale. We present them here as an early attempt at a survey. Please submit sources you've find relevant at the bottom of the page.

We exclude live deployments, which are listed in the org map.

90 papers.

Query syntax

Examples

socialone word, anywhere in an entry
"collective intelligence"that phrase, in any case
market OR economyeither word
safety NOT languagethe first without the second
institution:universityone field only
(social OR systemic) AND riskbrackets group

Fields

title:the paper's title
creator:any author's name
summary:the summary
tag:any of its tags
institution:any author's institution
date:the publication date
system:its system type
mix:its participant mix
focus:its focus area
threat:its threat models
claim:its claim type

Two words with no operator between them mean both. A term matches part of a word, so agent finds “agentic”. Your browser runs the query. It reaches no server.

Rows
Columns
Focus Area \ Threat ModelGradual DisempowermentSystemic InstabilityInequalityCollective SuperintelligencePartially Observable SystemsPower ConcentrationOutdated ModelsEmergent GoalsNot tagged
Monitoring
Steering
Simulation
Redesign
Design
Not tagged
Papers per pair, scaled to this view (max 34). Fewer is darker.01–910–1718–2627–34

All 90 papers

TitleCreatorsDateTags
What Do We Tell the Humans?Shoshannah TekofskyLive Systems
Generative AI as Economic AgentsNicole Immorlica; Brendan Lucier; Aleksandrs Slivkins2024
Causal Emergence 2.0: Quantifying Emergent ComplexityErik Hoel2025Collective Intelligence
Use of bot and content flags to limit the spread of misinformation among social networks: a behavior and attitude surveyCandice Lanius2021
Emergence of Scale-Free Networks in Social Interactions among Large Language ModelsGiordano De Marzo; Luciano Pietronero; David Garcia2023Collective Intelligence
Two types of AI existential risk: decisive and accumulativeFoundations & Taxonomies
Prompt Infection: LLM-to-LLM Prompt Injection within Multi-Agent SystemsDonghyun Lee; Mo Tiwari; Brian Miranda2024Security & Adversarial Robustness
How to Model AI Agents as Personas?: Applying the Persona Ecosystem Playground to 41,300 Posts on Moltbook for Behavioral InsightsDanial Amin; Joni Salminen; Bernard J. Jansen2026Persona Extraction; Datasets
System-level safety evaluations
OASIS: Open Agent Social Interaction Simulations with One Million AgentsZiyi Yang; Zaibin Zhang; Zirui Zheng; Yuxian Jiang; Philip Torr; Jing Shao2024Simulation and Digital Twins
Page 1 of 9

Transfer

Methods that transfer.

Though these sources study human systems without AI agents, they provide valuable background in humanistic large system settings, and may transfer to agentic systems as well.

15 papers.

Query syntax

Examples

socialone word, anywhere in an entry
"collective intelligence"that phrase, in any case
market OR economyeither word
safety NOT languagethe first without the second
institution:universityone field only
(social OR systemic) AND riskbrackets group

Fields

title:the paper's title
creator:any author's name
summary:the summary
tag:any of its tags
institution:any author's institution
date:the publication date
system:its system type
mix:its participant mix
focus:its focus area
threat:its threat models
claim:its claim type

Two words with no operator between them mean both. A term matches part of a word, so agent finds “agentic”. Your browser runs the query. It reaches no server.

Rows
Columns
Focus Area \ Threat ModelGradual DisempowermentSystemic InstabilityInequalityCollective SuperintelligencePartially Observable SystemsPower ConcentrationOutdated ModelsEmergent GoalsNot tagged
Monitoring
Steering
Simulation
Redesign
Design
Not tagged
Papers per pair, scaled to this view (max 5). Fewer is darker.01–2345

Contribute

Submit source.

Please submit relevant papers. We are looking to grow the list to a thorough coverage.