Artificial intelligence is rapidly becoming part of ordinary business operations. Companies are using AI to draft documents, communicate with customers, analyze data, create advertising, screen applicants, write software, and—increasingly—to perform tasks autonomously through AI agents.
That raises an increasingly important insurance question: What happens when the AI gets it wrong?
The answer may be more complicated than simply purchasing “AI insurance.” An AI-related loss can potentially implicate several traditional lines of insurance coverage, including cyber, errors and omissions, commercial general liability, crime, employment practices liability, and directors and officers insurance. At the same time, insurers are beginning to introduce policy language specifically addressing AI risks.
For policyholders, the important question is therefore not simply whether a policy says anything about artificial intelligence. It is what happened, what loss or claim resulted, and what coverage the policy provides for that loss.
The Rise of “Silent AI” Coverage
The insurance industry has seen this problem before.
Before cyber insurance became commonplace, cyber-related losses often landed in policies that were written long before anyone contemplated ransomware, data breaches, or sophisticated computer fraud. Insurers and policyholders spent years litigating whether those losses were covered under property, crime, general liability, and other traditional policies.
AI appears to be following a similar path.
The industry has begun using the term “silent AI” to describe potential coverage for AI-related risks under insurance policies that neither expressly cover nor exclude artificial intelligence. Many existing policies were simply not drafted with generative or autonomous AI in mind.
That does not necessarily mean there is no coverage. Insurance coverage generally turns on the language of the policy and the nature of the claim—not simply on the technology that happened to cause the loss.
Consider several increasingly plausible scenarios.
An AI Tool Makes a Professional Mistake
Suppose an accounting, consulting, engineering, technology, or other professional-services firm uses an AI tool in performing work for a client. The AI produces erroneous information, an employee fails to catch the error, and the client suffers a substantial financial loss.
The resulting claim may implicate the firm’s professional liability or errors and omissions coverage.
From a coverage standpoint, an important question may be whether the alleged wrongful act is the firm’s provision of professional services—not whether a human being or an AI system produced the erroneous information.
That distinction could become increasingly important as insurers begin adding AI-specific exclusions, limitations, or endorsements to professional liability policies.
An AI Agent Sends Money to the Wrong Place
AI is also becoming increasingly autonomous. Instead of merely producing text for a human to review, AI agents can potentially communicate with vendors, access systems, execute transactions, and initiate payments.
Imagine an AI agent responding to fraudulent instructions and transferring hundreds of thousands of dollars to a criminal.
That loss could raise issues under commercial crime, computer fraud, funds-transfer fraud, social-engineering, and cyber coverage.
Those disputes are already complicated when humans are deceived into transferring money. Adding an autonomous AI agent to the chain of events creates additional questions concerning causation, authorization, and the precise mechanism of the loss.
AI-Generated Content Creates Liability
Businesses increasingly use generative AI to create advertising copy, photographs, video, music, software, and other content.
What if that content allegedly infringes someone’s copyright, uses a person’s likeness without permission, or defames a competitor?
Depending on the allegations and policy language, the resulting claim could implicate commercial general liability, media liability, intellectual-property coverage, or technology E&O coverage.
Again, the coverage analysis should begin with the claim actually being asserted and the language of the applicable policy.
AI Makes an Employment Decision
Employers are increasingly using automated systems in hiring, performance management, scheduling, and other employment decisions.
If an applicant or employee alleges that an AI-assisted system discriminated against them, the resulting claim could potentially implicate employment practices liability insurance (EPLI).
The presence of AI does not necessarily change the fundamental nature of the underlying claim. An allegation of employment discrimination remains an allegation of employment discrimination even if software played a role in the decision.
But AI-specific exclusions or definitions could change that analysis, making it increasingly important for businesses to review new policy forms and endorsements at renewal.
Management Gets AI Wrong
AI-related claims may also reach the boardroom.
Public companies have already faced scrutiny concerning representations about their use and capabilities of artificial intelligence. Companies may also face allegations that directors or officers failed to supervise AI deployment, adequately disclose AI-related risks, implement appropriate controls, or respond appropriately after an AI-related incident.
Those claims may implicate directors and officers liability insurance.
The issue becomes particularly important because an AI event can potentially implicate several policies simultaneously. A single incident might generate a cyber claim, customer litigation, regulatory proceedings, and claims against management.
The Insurance Market Is Already Responding
Businesses should not assume that today’s coverage will look like yesterday’s coverage.
The insurance industry is actively developing AI-specific products, endorsements, sublimits, and exclusions. Optional exclusions addressing generative AI have already been developed for commercial general liability policies, while other insurers are offering affirmative coverage for particular AI risks.
This means policyholders should pay particular attention to changes made at renewal.
A policy that previously said nothing about artificial intelligence may suddenly contain an endorsement addressing generative AI, automated decision-making, algorithms, or related technology. Even where no provision is labeled an “AI exclusion,” changes to definitions, insuring agreements, exclusions, or sublimits may materially affect coverage.
Don’t Start With the Robot
When an AI-related claim occurs, insurers and policyholders may be tempted to begin with the novelty of the technology.
That may be the wrong place to start.
Insurance coverage analysis should begin with the same questions that apply to any other claim:
What happened? What damages or loss resulted? What claim is being asserted against the policyholder? Which insuring agreements potentially apply? And does the insurer have policy language that actually excludes the loss?
If an AI tool causes a professional error, the claim may still be a professional liability claim.
If AI-generated content causes an advertising injury, the claim may still involve advertising injury.
If an AI-assisted attack steals money, the loss may still implicate crime or cyber coverage.
The fact that artificial intelligence appears somewhere in the causal chain should not, standing alone, determine whether insurance coverage exists.
What Businesses Should Do Now
Businesses using AI should consider reviewing their insurance programs before a significant loss occurs.
That review should include identifying how AI is actually being used within the organization and comparing those uses against existing cyber, professional liability, general liability, crime, EPLI, D&O, and other potentially applicable policies.
Policyholders should also carefully compare renewal policies against expiring coverage. As insurers begin addressing AI more expressly, seemingly modest changes to endorsements or definitions may create significant new gaps.
Finally, businesses should resist assuming that a loss involving artificial intelligence is automatically uninsured—or that an insurer’s characterization of something as an “AI loss” resolves the coverage question.
Artificial intelligence may be new.
The fundamental rules governing insurance coverage are not.
And as businesses increasingly entrust important decisions to machines, figuring out who insures the robot may become one of the more important insurance coverage questions of the next several years.
Jim Guse • Partner
Barker Martin, P.S.
Direct (971) 277-4711
www.barkermartin.com
