Polarization by Design - AI Persuasion Simulator

Simulation Speed

Fast 10x Slow

Visualization Style

Template Scenarios

Select a scenario to load preset configurations that demonstrate different polarization dynamics from the paper.

Scenario Mode

AI Technology Parameters

Persuasion Cost (lower = more powerful AI)

1.0

Persuasion Reach (% of population)

30%

Message Intensity

0.50

Population Parameters

Initial Distribution (spread)

0.25

Resistance to Change

0.50

Population Size

500

Competing Elites Settings

Power Transition Rate

0.05

Elite B Strength

0.70

Opinion Distribution

Time Step

0.00

Polarization Index

Ruling Elite

Elite A Target (Right)

Elite B Target (Left)

Population Distribution

Semi-Lock Region

Polarization Over Time

Understanding the Polarization Index

The Polarization Index (0.0 - 1.0) measures how divided the population's opinions are. It combines three factors from the paper's model:

Variance: How spread out opinions are from the mean (higher = more dispersed)
Bimodality: Proportion of population at extremes (<0.4 or >0.6) vs. center
Center Emptiness: How few people remain moderate (0.4-0.6 range)

Reading the Graph: A rising line shows increasing polarization - the elite's strategy is working. Flat lines indicate equilibrium. In competing mode, oscillations reflect power transitions between elites. The paper argues that as AI persuasion costs drop, the equilibrium polarization level rises - making extreme division the optimal governance strategy.

Current Dynamics

The simulation models how AI-driven persuasion technology affects opinion distributions. Lower persuasion costs enable elites to more effectively push populations toward extreme positions. Start the simulation to observe the "polarization pull" effect described in the paper.