The insurance industry is drawing a line between AI hype and hard liability. A new report from the RAND Corporation finds that insurers are increasingly wary of covering corporate AI harms, with some carriers already introducing exclusions for generative AI failures. The mismatch between rapid enterprise AI adoption and a fragmented insurance market leaves companies exposed to financial liability from deepfakes, hallucinations, privacy violations, and discriminatory outputs that do not fit neatly into existing policies.
RAND says its report was motivated by the gap between the speed of AI deployment and the slow, uncertain response from insurers, regulators, and brokers. "AI-related harms are already emerging, including incorrect or misleading outputs, deepfakes, privacy violations, intellectual property disputes, fraud, product defects, and discriminatory decisions," the report states. "These harms create demand for insurance but do not fit neatly within existing insurance lines."
If AI becomes uninsurable, the consequences extend beyond the insurance sector. Corporate customers may delay AI projects to meet fiduciary obligations, and the AI industry itself would face pressure to moderate its ambitions and sales targets.
Carriers begin adding AI exclusions
W. R. Berkley has already introduced exclusions in its directors and officers (D&O), errors and omissions (E&O), and fiduciary liability products to exclude coverage for "any actual or alleged use, deployment, or development of Artificial Intelligence." During the company's Q4 2025 earnings call, CEO W. Robert Berkley said underwriters must understand "the impact that [new technologies like AI are] having on our insureds, what it means for risk, and our ability to fully understand that risk so we can control it, select it, and price for it."
In January 2026, Verisk/ISO - whose standardized forms appear in more than 80 percent of US property and casualty policies - introduced optional language carriers can adopt to exclude bodily injury, property damage, and other harms arising from generative AI. The move signals a structural shift in how the industry approaches machine-made mishaps.
What the incident data shows
The Artificial Intelligence Incident Database (AIIDB) lists 713 incidents drawn from more than 6,000 reports, covering AI use beyond chatbots. The categorical breakdown includes: misinformation and manipulation (586), deepfakes and synthetic media (346), deepfake-enabled misinformation (333), hallucination and factual error (215), harmful content (92), agentic and autonomous failure (84), privacy and data leak (58), bias and discrimination (47), copyright and IP (20), and other categories. Many incidents span multiple classifications.
Beyond incident reports, approximately 250 US lawsuits related to AI are working through courts. These cases largely focus on copyright and IP but also touch on privacy and surveillance, fraud and deception, negligence and product liability, discrimination and civil rights, and contracts and trade secrets. Dozens of state laws now cover AI-generated intimate images, AI-generated child sexual abuse material, automated decision-making, and AI in political ads - all creating compliance risks for businesses deploying AI.
Not all insurers are backing away
Some carriers see the uncertainty as an opportunity. Coverage gaps are being filled by new and existing companies that believe they can calculate the risk. The market for AI for Insurance is fragmenting into specialists who underwrite what generalists exclude. RAND argues that a common taxonomy to track AI incidents and claims is essential for brokers, carriers, and reinsurers to price these risks consistently.
The think tank also wants state regulators to push for an AI Coverage Notice so policyholders understand exactly what is covered and what is not. RAND expects AI insurance to move beyond a specialty product as the risks become better understood - but that will take time. Effective AI Risk Management depends on data the industry does not yet have.
Why this matters for insurance professionals
Underwriters and brokers face a product design problem with real urgency. Exclusions are already appearing in standardized forms from Verisk/ISO, and the 250 active AI lawsuits mean claims are not hypothetical. The pressure to clarify coverage language will come from both corporate clients demanding protection and regulators demanding transparency. Professionals who understand the incident taxonomy and the evolving legal landscape will be positioned to write policies that address specific AI harms - or to advise clients on where coverage is disappearing.
Your membership also unlocks:
