10,000 People AI Evolution Experiment Program

10,000 People AI Evolution Experiment Program
Organizer：AFE US Company
Experiment Code：Quantiveon Evolution

1.Experiment Background
Core Proposition: To verify the adaptability of AI systems in human-led environments
Scientific Value: Establishing, for the first time, a bidirectional co-evolution model between human emotional feedback and AI strategy adjustment
Business Objective: To provide strategy optimization support for a future $1 billion AI fund
Social Significance: Achieving financial democratization through AI technology

2.Experiment Design
(1) Technical Architecture: Dual engine driven
AI System: FMAI 3.0 (integrating reinforcement learning and game theory modules)
Human Intervention: Risk Control Committee composed of two CFA charter holders
Dynamic Takeover Threshold：
∣σpred−σhist∣/σhist>20%|\sigma_{pred} – \sigma_{hist}| / \sigma_{hist} > 20\%
(2) Participation Mechanism ：
TIER Capital Range Quota Target Annualized Return
A $600,000 – $1,000,000 1,000pax ≥500%
B $200,000 – $600,000 3,000pax ≥400%
C $100 – $200,000 6,000pax ≥300%
(3) Risk Control Mechanism：
Loss Compensation Reserve Fund：$100 million
Maximum Drawdown Circuit Breaker: System suspension triggered if single-day loss reaches 5%

3.Execution Process
(1) Experiment Duration: Total 17 months (August 1, 2025 – December 31, 2026)
(2) Volunteer Application Period: July 1, 2025 – December 31, 2025
(3)Individual Volunteer Participation Period: Maximum of 1 year

4.Volunteer Recruitment Methods
Media Advertising
Volunteer Referral System
5.Data Collection Methods

Sentiment Indicator: Quantifying the “Greed Index” through trading frequency and cancellation rates
AI Evolution Verification: Comparing changes in Sharpe ratios between Q3 and Q4
Incident Records: Documenting AI strategy failures during “Black Swan” events

AFE AI 3.0 Ten Thousand People AI Evolution Experiment Working Group
June 6, 2025