Insurers Are Adding AI Exclusions to Policies. Courts May Not Enforce Them.
Insurance carriers are rapidly introducing exclusions designed to eliminate coverage for artificial intelligence-related claims, but the language is so broad it may not survive legal scrutiny. The exclusions could render policies worthless for businesses that rely on AI, and courts have established doctrines that work against such sweeping carve-outs.
Berkley, Hamilton Insurance Group, and the Insurance Services Office have all introduced AI exclusions in recent months. Some are written to exclude any claim "based upon, arising out of, or attributable to" AI use, development, or disclosure. Berkley's definition of artificial intelligence is expansive enough to cover machine-learning systems that have been in use for years before the recent generative AI boom.
The Litigation Driving the Exclusions
Insurers are responding to a widening range of AI-related lawsuits. Plaintiffs have filed claims involving copyright infringement from training data, product liability from AI system failures, privacy violations from data scraping, antitrust allegations, algorithmic discrimination, and securities fraud over misrepresented AI capabilities.
A copyright case against Anthropic reportedly settled for $1.5 billion. A discrimination class action in federal court alleged that Workday's AI systems produced biased hiring outcomes. These cases span federal and state courts across the country.
Why Broad Exclusions May Not Hold
Exclusions are read narrowly. Insurance law favors coverage. Courts consistently interpret exclusionary language narrowly and against the insurer. An exclusion must plainly and clearly bar a claim-not merely touch on a topic somewhere in the complaint.
If a lawsuit includes both AI-related allegations and non-AI wrongdoing, the insurer typically must defend the entire case. A claim involving AI-related conduct plus traditional mismanagement, contractual disputes, or operational errors may still trigger coverage for the non-AI elements.
Exclusions cannot eliminate the policy's purpose. Courts have rejected exclusions that are so broad they "swallow" the coverage the policy promised. If an AI exclusion would eliminate coverage for nearly all of a company's operations-particularly for technology firms or data-driven businesses-a court may find the exclusion unreasonable.
As AI becomes embedded in marketing, customer service, logistics, hiring, and financial forecasting, almost any business claim could touch on AI tangentially. Enforcing the broadest versions of these exclusions would render insurance policies useless for entire industries.
Earlier Policies May Still Provide Coverage
Many companies maintain multiyear occurrence-based liability policies that predate AI exclusions. These older policies contain no language limiting AI claims. When an insurer adds an AI exclusion to new policies, courts sometimes view that addition as evidence that earlier policies were broader and did not exclude AI-related claims.
A company facing AI litigation should examine policies issued in prior years. If the alleged wrongful acts span multiple periods, coverage under earlier policies may be available.
What Policyholders Should Do Now
Review new and renewal policies carefully. AI exclusions are evolving rapidly, and policyholders should work with brokers to negotiate policies without such exclusions or to narrow their definitions.
Be precise in insurance applications about AI capabilities, governance, and controls. Vague or overstated representations about AI will complicate coverage disputes later.
Do not assume an AI exclusion automatically bars coverage. Mixed allegations, independent causes of injury, and established principles of narrow construction may preserve defense and indemnity rights.
For businesses deeply integrated with AI, consider arguing that overly broad exclusions conflict with the basic purpose of the policy. This argument may be particularly strong for technology companies.
Experienced coverage counsel should evaluate these issues at policy placement, renewal, and when assessing potential claims. The human element remains crucial in navigating AI insurance disputes.
For more on AI for Insurance and AI for Legal Professionals, see our dedicated resources.
Your membership also unlocks:
