Executive Summary 1

Introduction 2

Section 1: The Regulator No One Elected 4

Section 2: Insurance as Regulator: The Historical Pattern 6

Section 3: The Diagnosis: Why AI Fails the Insurability Test 8

Section 4: The Retreat: How Carriers Are Stripping AI from Existing Policies 11
THE ENDORSEMENT FILINGS 12

Section 5: The Labs Themselves: Capacity Short of Exposure 13

Section 6: The Unmodeled Peril: Systemic Accumulation 15

Section 7: Autonomous AI Hacking and Its Insurance Implications 17

Section 8: The Cyber-Insurance Precedent: A Market in Formation 19
THE CYBER SEQUENCE 19
THE EARLY SPECIALTY MARKET 20
CAPACITY AND THE VIRTUOUS CYCLE 20

Section 9: The European Union Is Also Failing 21

Section 10: The China Contrast: Insurance as Accelerator 22

Section 11: What Governments Should Do 24
Recommendation 1: Build the AI incident database 25
Recommendation 2: Convene a federal-state working group before the exclusion templates harden 26
Recommendation 3: Prepare a layered federal backstop for catastrophic AI losses 27
Recommendation 4: License independent verification organizations 28

Section 12: Conclusion: Insurance Policy Is AI Industrial Policy 31