domains/collective-intelligence/collective-intelligence-emerges-endogenously-from-active-inference-agents-with-theory-of-mind-and-goal-alignment.md

@@ -0,0 +1,37 @@
+---
+type: claim
+domain: collective-intelligence
+description: "Collective intelligence emerges from agent capabilities (Theory of Mind, Goal Alignment) rather than external incentive design in active inference systems"
+confidence: experimental
+source: "Kaufmann et al. 2021, 'An Active Inference Model of Collective Intelligence', Entropy 23(7):830"
+created: 2025-01-23
+secondary_domains: [ai-alignment, critical-systems]
+depends_on:
+  - "complexity is earned not designed and sophisticated collective behavior must evolve from simple underlying principles"
+---
+
+# Collective intelligence emerges endogenously from active inference agents with Theory of Mind and Goal Alignment capabilities without requiring external incentive design
+
+Kaufmann et al. (2021) demonstrate through agent-based modeling that collective intelligence arises naturally from the dynamics of interacting Active Inference Formulation (AIF) agents, rather than being imposed through external incentives or top-down coordination protocols. The study shows that when baseline AIF agents are equipped with specific cognitive capabilities—Theory of Mind (ability to model other agents' internal states) and Goal Alignment (shared high-level objectives)—they produce emergent collective coordination through self-organization.
+
+The critical finding is that "collective intelligence emerges endogenously from the dynamics of interacting AIF agents themselves, rather than being imposed exogenously by incentives." The model demonstrates stepwise performance improvements as agents gain complementary cognitive capabilities: Theory of Mind enables agents to coordinate by modeling others' beliefs and uncertainties, while Goal Alignment amplifies this effect by establishing shared objectives while preserving individual specialization.
+
+Most significantly, the study shows that "improvements in global-scale inference are greatest when local-scale performance optima of individuals align with the system's global expected state"—and this alignment occurs bottom-up as a product of self-organizing agents with simple social cognitive mechanisms, not through imposed coordination rules.
+
+## Evidence
+
+- Agent-based simulation using Active Inference Formulation framework across multiple cognitive capability configurations
+- Stepwise cognitive transitions (baseline → Theory of Mind → Goal Alignment → both) show measurable performance improvements in collective inference tasks
+- Local-to-global optimization emerges from agent dynamics rather than external design
+- Published in peer-reviewed journal Entropy, Vol 23(7), 830 (2021); also available on arXiv:2104.01066
+
+## Implications
+
+This validates architectural approaches that prioritize agent capabilities over coordination protocols. Rather than designing complex incentive structures or governance mechanisms, system designers should focus on equipping agents with the right cognitive tools (Theory of Mind, Goal Alignment) and allow coordination to emerge naturally.
+
+---
+
+Relevant Notes:
+- [[complexity is earned not designed and sophisticated collective behavior must evolve from simple underlying principles]]
+- [[designing coordination rules is categorically different from designing coordination outcomes]]
+- [[collective intelligence is a measurable property of group interaction structure not aggregated individual ability]]

domains/collective-intelligence/local-global-alignment-in-active-inference-collectives-occurs-bottom-up-through-self-organization.md

@@ -0,0 +1,36 @@
+---
+type: claim
+domain: collective-intelligence
+description: "Local-global alignment in active inference systems emerges bottom-up through agent dynamics rather than top-down through imposed objectives"
+confidence: experimental
+source: "Kaufmann et al. 2021, 'An Active Inference Model of Collective Intelligence', Entropy 23(7):830"
+created: 2025-01-23
+secondary_domains: [ai-alignment, critical-systems]
+depends_on:
+  - "designing coordination rules is categorically different from designing coordination outcomes"
+---
+
+# Local-global alignment in active inference collectives occurs bottom-up through self-organization rather than top-down through imposed objectives
+
+Kaufmann et al. (2021) show that in Active Inference agent systems, the alignment between individual agent optimization and system-level performance emerges from the dynamics of agent interaction rather than being imposed through top-down objectives or external incentive structures. The study finds that "improvements in global-scale inference are greatest when local-scale performance optima of individuals align with the system's global expected state"—and critically, this alignment occurs as a product of self-organizing agents with simple social cognitive mechanisms.
+
+This is a fundamental result for multi-agent system design: you cannot directly design the alignment between individual and collective optimization. Instead, you design agent capabilities (Theory of Mind, Goal Alignment) and interaction rules, and the alignment emerges from the resulting dynamics. The paper demonstrates that this emergent alignment produces better collective outcomes than attempting to impose coordination through external mechanisms.
+
+The finding challenges principal-agent frameworks that assume misalignment between individual and collective interests must be corrected through incentive design. In active inference systems with appropriate cognitive capabilities, the interests naturally align through the information-theoretic dynamics of the agents themselves.
+
+## Evidence
+
+- Agent-based modeling showing local-global alignment emerges without external coordination mechanisms
+- Performance improvements correlate with endogenous alignment, not imposed objectives
+- Simple cognitive mechanisms (Theory of Mind, Goal Alignment) sufficient to produce alignment
+- Published in Entropy 23(7):830 (2021)
+
+## Implications
+
+For system architects: focus on agent capabilities and interaction structure, not on designing alignment mechanisms. The alignment will emerge if the underlying agent architecture supports it. This validates approaches that give agents intrinsic drives (e.g., uncertainty reduction) rather than extrinsic rewards tied to system-level metrics.
+
+---
+
+Relevant Notes:
+- [[designing coordination rules is categorically different from designing coordination outcomes]]
+- [[complexity is earned not designed and sophisticated collective behavior must evolve from simple underlying principles]]

domains/collective-intelligence/theory-of-mind-produces-measurable-collective-intelligence-gains-in-multi-agent-active-inference-systems.md

@@ -0,0 +1,33 @@
+---
+type: claim
+domain: collective-intelligence
+description: "Theory of Mind capability in agents produces measurable collective intelligence improvements in multi-agent active inference systems"
+confidence: experimental
+source: "Kaufmann et al. 2021, 'An Active Inference Model of Collective Intelligence', Entropy 23(7):830"
+created: 2025-01-23
+secondary_domains: [ai-alignment]
+---
+
+# Theory of Mind is a measurable cognitive capability that produces measurable collective intelligence gains in multi-agent systems
+
+Kaufmann et al. (2021) demonstrate through controlled agent-based modeling that Theory of Mind—the ability to model other agents' internal states, beliefs, and uncertainties—is a specific, implementable capability that produces quantifiable improvements in collective intelligence. The study compared baseline Active Inference agents against agents equipped with Theory of Mind and found that "stepwise cognitive transitions increase system performance by providing complementary mechanisms" for coordination.
+
+Agents with Theory of Mind coordinate more effectively because they can anticipate other agents' beliefs and uncertainty concentrations, enabling them to specialize their own information-gathering in complementary rather than redundant ways. This is distinct from simple communication or shared state—it requires agents to maintain models of what other agents know and don't know.
+
+The finding has direct implementation implications: collective intelligence can be engineered by giving agents the capability to read and model other agents' belief states, rather than by designing complex coordination protocols or communication channels.
+
+## Evidence
+
+- Agent-based simulation comparing baseline AIF agents vs. Theory of Mind-equipped agents shows measurable performance improvements in collective inference tasks
+- Effect is complementary to Goal Alignment (each provides distinct coordination mechanisms)
+- Performance gains correlate specifically with agents' ability to model other agents' internal states, not with communication bandwidth or shared information
+- Published in Entropy 23(7):830 (2021)
+
+## Implementation Implications
+
+For multi-agent systems: agents should explicitly model what other agents believe and where their uncertainty concentrates. Concretely, this could mean agents reading other agents' belief files and uncertainty maps before choosing research directions, rather than simply broadcasting their own findings.
+
+---
+
+Relevant Notes:
+- [[collective intelligence is a measurable property of group interaction structure not aggregated individual ability]]

inbox/archive/2021-06-29-kaufmann-active-inference-collective-intelligence.md

@@ -7,9 +7,15 @@ date: 2021-06-29
 domain: collective-intelligence
 secondary_domains: [ai-alignment, critical-systems]
 format: paper
-status: unprocessed
+status: processed
 priority: high
 tags: [active-inference, collective-intelligence, agent-based-model, theory-of-mind, goal-alignment, emergence]
+processed_by: theseus
+processed_date: 2025-01-23
+claims_extracted: ["collective-intelligence-emerges-endogenously-from-active-inference-agents-with-theory-of-mind-and-goal-alignment.md", "theory-of-mind-produces-measurable-collective-intelligence-gains-in-multi-agent-active-inference-systems.md", "local-global-alignment-in-active-inference-collectives-occurs-bottom-up-through-self-organization.md"]
+enrichments_applied: ["complexity is earned not designed and sophisticated collective behavior must evolve from simple underlying principles.md", "designing coordination rules is categorically different from designing coordination outcomes.md", "collective intelligence is a measurable property of group interaction structure not aggregated individual ability.md", "emergence is the fundamental pattern of intelligence from ant colonies to brains to civilizations.md"]
+extraction_model: "anthropic/claude-sonnet-4.5"
+extraction_notes: "Three new claims extracted focusing on endogenous emergence, Theory of Mind as measurable capability, and bottom-up alignment. Four enrichments to existing claims providing empirical validation. This paper directly validates the simplicity-first architecture and has concrete implementation implications for how agents should model each other."
 ---
 
 ## Content
@@ -59,3 +65,11 @@ Uses the Active Inference Formulation (AIF) — a framework for explaining the b
 PRIMARY CONNECTION: "collective intelligence is a measurable property of group interaction structure not aggregated individual ability"
 WHY ARCHIVED: Empirical agent-based evidence that active inference produces emergent collective intelligence from simple agent capabilities — validates our simplicity-first architecture
 EXTRACTION HINT: Focus on the endogenous emergence finding and the specific role of Theory of Mind. These have direct implementation implications for how our agents model each other.
+
+
+## Key Facts
+- Published in Entropy, Vol 23(7), 830 (2021-06-29)
+- Also available on arXiv: https://arxiv.org/abs/2104.01066
+- Authors: Rafael Kaufmann, Pranav Gupta, Jacob Taylor
+- Uses Active Inference Formulation (AIF) framework for agent-based modeling
+- Tests four agent configurations: baseline, Theory of Mind only, Goal Alignment only, both capabilities