Insurance Liability in the Age of AI Transformation

As AI advances at a rapid pace, the fundamental tort law concepts that undergird insurance and liability regimes are being put to the test. The landscape is evolving fast. As autonomous vehicles (AV) become more widespread, the demand for traditional auto insurance is shrinking, while an explosive growth in insurance demand is projected for AI risks related to agentic AI and cybersecurity. Underwriting a wide range of liabilities will require innovative approaches to risk management.

As AVs become more widespread, we will likely see a significant reduction in the frequency and nature of accidents. According to a McKinsey report, as much as 90% of accidents could be eliminated with a mass adoption of AVs, dramatically diminishing the need for traditional auto insurance. Furthermore, the focus of auto insurance will shift from negligence to product liability. Traditional auto insurance assesses driver negligence. However, how can you assign fault to an autonomous system in an AV? Liability will shift towards manufacturers and developers of AVs in the form of product liability. Large AV manufacturers may choose to self-insure and pay out compensation to victims directly since they may have the resources to do so, further decreasing the need for insurance in the auto industry.

While the traditional auto insurance industry will face decline, there will be a surge of insurance demand for agentic AI and cybersecurity-related risks. Agentic AI is a system that can make independent decisions and execute plans without direct human oversight. While nascent at this stage, it will increasingly automate routine tasks. Agentic AI brings new, complex risks from misuse and misalignment that traditional policies cannot capture. Advanced AI systems introduce new vulnerabilities, such as data poisoning, adversarial attacks, and AI-driven fraud. Insurers must develop specialized products that address dynamic threats and deliver real-time, customized risk assessments.

Unlike the auto industry, where large companies often have the resources to pay out substantial damages, many AI developers are start-ups with limited funds. Insurance mandates may help guarantee a minimum level of victim compensation while also offering clear financial incentives for AI companies to implement strong risk management practices. This approach could encourage greater responsibility and prudence in the AI sector.