Entity Relationship Graph

This interactive graph shows the relationships between entities in the knowledge base. Nodes are color-coded by entity type, and edges represent relationships defined in the entity data.

How to Use

Pan: Click and drag on the background
Zoom: Scroll or pinch
Select: Click on a node to see details
Move nodes: Drag individual nodes to rearrange

International Coordination

Societal Trust

Epistemic Health

Information Authenticity

AI Control Concentration

Human Agency

Economic Stability

Human Expertise

Human Oversight Quality

Alignment Robustness

Safety-Capability Gap

Interpretability Coverage

Regulatory Capacity

Institutional Quality

Reality Coherence

Preference Authenticity

Racing Intensity

Safety Culture Strength

Coordination Capacity

Biological Threat Exposure

Cyber Threat Exposure

Societal Resilience

Alignment Progress

Safety Research

Lab Behavior

Public Opinion

Expert Opinion

Economic & Labor

AI Capabilities

Compute & Hardware

Geopolitics

Structural Indicators

Existential Catastrophe

Long-term Trajectory

AI Takeover

Human-Caused Catastrophe

Long-term Lock-in

Misalignment Potential

Misuse Potential

AI Capabilities

AI Uses

AI Ownership

Civilizational Competence

Transition Turbulence

Agentic AI

Autonomous Coding

Large Language Models

Long-Horizon Autonomous Tasks

Persuasion and Social Manipulation

Reasoning and Planning

RLHF

Scientific Research Capabilities

Self-Improvement and Recursive Enhancement

Situational Awareness

Tool Use and Computer Use

Corporate Influence

Field Building and Community

AI Governance and Policy

Research Agendas

Technical AI Safety Research

AI Misuse

Dual-Use Technology

Fast Takeoff

Superintelligence

Content Moderation

AGI Race

Capability Evaluations

Existential Risk

Adversarial Robustness

Natural Abstractions

AI Benchmarking

Transformative AI

Scaling Laws

AI Timelines

Data Constraints

Deep Learning Revolution Era

Mainstream Era

The MIRI Era

AI Safety Summit (Bletchley Park)

Lock-in Mechanisms Model

AI Risk Portfolio Analysis

Worldview-Intervention Mapping

Intervention Timing Windows

Deceptive Alignment Decomposition Model

Carlsmith's Six-Premise Argument

Mesa-Optimization Risk Analysis

Goal Misgeneralization Probability Model

Reward Hacking Taxonomy and Severity Model

Power-Seeking Emergence Conditions Model

Instrumental Convergence Framework

Scheming Likelihood Assessment

Corrigibility Failure Pathways

Bioweapons Attack Chain Model

AI Uplift Assessment Model

AI-Bioweapons Timeline Model

Racing Dynamics Impact Model

Multipolar Trap Dynamics Model

Flash Dynamics Threshold Model

Expertise Atrophy Progression Model

Economic Disruption Impact Model

AI Proliferation Risk Model

Winner-Take-All Concentration Model

Cyber Offense-Defense Balance Model

Autonomous Cyber Attack Timeline

Autonomous Weapons Escalation Model

LAWS Proliferation Model

Disinformation Detection Arms Race Model

Electoral Impact Assessment Model

AI Surveillance and Regime Durability Model

Surveillance Chilling Effects Model

Deepfakes Authentication Crisis Model

Trust Cascade Failure Model

Sycophancy Feedback Loop Model

Authentication Collapse Timeline Model

Expertise Atrophy Cascade Model

Epistemic Collapse Threshold Model

Reality Fragmentation Network Model

Racing Dynamics Game Theory Model

Multipolar Trap Coordination Model

Winner-Take-All Market Dynamics Model

Concentration of Power Systems Model

Lock-in Irreversibility Model

Economic Disruption Structural Model

AI Capability Proliferation Model

Risk Activation Timeline Model

Capability Threshold Model

Warning Signs Model

Authoritarian Tools Diffusion Model

Consensus Manufacturing Dynamics Model

Irreversibility Threshold Model

Preference Manipulation Drift Model

Trust Erosion Dynamics Model

Automation Bias Cascade Model

Cyber Psychosis Cascade Model

Fraud Sophistication Curve Model

Intervention Effectiveness Matrix

Lab Incentives Model

Risk Interaction Matrix

Safety Research Value Model

Capabilities-to-Safety Pipeline Model

Compounding Risks Analysis Model

Defense in Depth Model

Institutional Adaptation Speed Model

International Coordination Game Model

Media-Policy Feedback Loop Model

Public Opinion Evolution Model

Risk Cascade Pathways Model

Risk Interaction Network Model

Safety-Capability Tradeoff Model

Safety Research Allocation Model

Entity Types

risk

risk-factor

safety-agenda

intervention

policy

capability

model

crux

concept

organization

Graph Stats

Nodes: 150

Edges: 295

Orphans: 38

Clusters: 21

React Flow

Most Connected Entities

Anthropic (32)
Racing Dynamics (31)
Alignment Robustness (22)
OpenAI (22)
Interpretability (20)

Largest Clusters

Anthropic (58 entities)
Racing Dynamics (57 entities)
Societal Trust (55 entities)
Concentration of Power (13 entities)
Compute Governance (12 entities)

Graph Analysis

Most Connected Entities

These entities have the most relationships:

Entity	Connections

| Anthropic | 32 | | Racing Dynamics | 31 | | Alignment Robustness | 22 | | OpenAI | 22 | | Interpretability | 20 | | Deceptive Alignment | 20 | | Compute Governance | 19 | | Concentration of Power | 18 | | Scalable Oversight | 17 | | Societal Trust | 15 |

Clusters Detected

The graph has 21 clusters (groups of interconnected entities):

Cluster around anthropic: 58 entities

Cluster around racing-dynamics: 57 entities

Cluster around societal-trust: 55 entities

Cluster around concentration-of-power: 13 entities

Cluster around compute-governance: 12 entities

Cluster around irreversibility: 12 entities

Cluster around self-improvement: 10 entities

Cluster around economic-disruption: 9 entities

Orphaned Entities

38 entities have no connections

misaligned-catastrophe
slow-takeoff-muddle
aligned-agi
multipolar-competition
pause-and-redirect
is-ai-xrisk-real
open-vs-closed
pause-debate
agi-timeline-debate
regulation-debate
interpretability-sufficient
scaling-debate
why-alignment-easy
case-against-xrisk
why-alignment-hard
case-for-xrisk
early-warnings
effectiveness-assessment
key-publications
post-incident-recovery
...and 18 more

About This Visualization

The graph uses label propagation for cluster detection, which groups entities that have strong interconnections. The centrality metric (number of connections) helps identify key hub entities.

Entity types use consistent colors:

🔴 Red: Risks (outcomes, pathways, amplifiers)
🟢 Green: Safety agendas and interventions
🔵 Blue: Policies and organizations
🟣 Purple: Capabilities and concepts
🟠 Amber: Models