INSURERS CAP PAYOUTS FOR AI AND LLMJACKING LOSSES

The insurance industry is moving to restrict payouts for damages stemming from AI deployment and LLMjacking—attacks involving the hijacking of large language models—according to documents reviewed by the Financial Times. QBE and Beazley, among other carriers, are implementing caps on cyber policy coverage for losses and regulatory penalties linked to AI systems. The shift reflects growing uncertainty around liability exposure as AI technology evolves faster than risk assessment frameworks. LLMjacking represents a nascent threat where attackers compromise AI models to generate malicious outputs or extract sensitive data. Insurers lack sufficient historical data to accurately price these risks, prompting them to establish financial guardrails. The moves signal insurers' cautiousness about underwriting AI-related claims while the technology and its regulatory landscape remain fluid. Organizations relying on AI systems may face reduced coverage options or higher premiums as the industry recalibrates its risk models.