Pennsylvania Settles with GEICO Over AI Claims Cancellation Without Notice
Pennsylvania Attorney General David Sunday reached a settlement with GEICO in May 2026 over the insurer's use of an AI tool that cancelled a customer's auto policy without adequate notice. The case centers on a Philadelphia policyholder whose coverage was terminated during GEICO's standard 60-day review for new customers, leaving her unknowingly uninsured.
GEICO's AI system flagged the applicant for additional underwriting and requested more information. When the company deemed the response insufficient, it cancelled the policy. The consumer continued driving under the assumption her coverage remained active.
What the Investigation Found
The Attorney General alleged GEICO engaged in unfair or confusing insurance practices. The issue wasn't AI use itself, but how the tool interacted with customer communication and decision-making processes. Officials expressed concern about the fairness and clarity of procedures, particularly for consumers selected for review by algorithmic systems.
The settlement signals regulatory attention on automated underwriting and decision-making in insurance, especially where such systems affect coverage continuity.
Settlement Requirements
GEICO must now adhere to Pennsylvania Insurance Department guidance on AI systems, based on the NAIC's AI Model Bulletin. The requirements include:
- Any AI-supported decision affecting a consumer must comply with all applicable state and federal insurance regulations
- Develop and maintain a formal AI governance program with executive oversight and risk management
- Implement processes to detect and mitigate AI biases to prevent unfair discrimination against protected classes
- Remain fully responsible for third-party AI systems and ensure they comply with Pennsylvania insurance law
- Disclose AI system use and provide detailed algorithmic documentation to regulators during examinations or investigations
What This Means for Insurers
Regulators are unlikely to challenge AI use in underwriting itself. They will scrutinize whether automated systems produce opaque outcomes, inadequate communication, or procedurally unfair results-especially in high-stakes decisions like cancellation.
Insurers should treat AI outputs as regulatory decisions, not technical artifacts. This means embedding compliance, legal review, and auditability into every stage of the underwriting process.
Strengthen customer communication controls across the board: clear notices, confirmation of deficiencies, and reasonable cure periods. Robust AI governance frameworks aligned with NAIC guidance ensure model decisions are explainable, documented, and subject to human oversight.
Learn more about AI for Insurance and AI for Legal to understand how these tools fit into broader compliance frameworks.
Your membership also unlocks:
