Artificial intelligence (AI) is no longer a hypothetical issue for community associations. As the technology becomes more accessible, board members are using AI tools with increasing frequency, often without fully appreciating the legal, practical, and insurance risks. Becker Shareholder Donna DiMaggio Berger recently addressed several risks associated with AI use by community associations. But the concern extends beyond whether a board member uses the technology appropriately. Associations must also consider whether a board member’s use of AI could jeopardize insurance coverage when a governance decision results in a claim.
In 2025, W. R. Berkley Corporation (Berkley), a major commercial-lines insurer, introduced an endorsement described as an “absolute” AI exclusion for certain directors and officers, errors and omissions, and fiduciary liability policies. The endorsement broadly targets claims connected to AI use, including the use of AI-generated content in decision-making.
The exclusion provides, in pertinent part:
The Insurer shall not be liable to make payment under this Coverage Part for Loss on account of any Claim made against any Insured based upon, arising out of, or attributable to:
(1) any actual or alleged use, deployment, or development of Artificial Intelligence by any person or entity, including but not limited to:
(a) the generation, creation, or dissemination of any content or communications using Artificial Intelligence;
* * *
(c) any Insured’s inadequate or deficient policies, practices, procedures, or training relating to Artificial Intelligence or failure to develop or implement any such policies, practices, procedures, or training;
(d) any Insured’s actual or alleged breach of any duty or legal obligation with respect to the . . . use . . . [or] deployment . . . of Artificial Intelligence;
* * *
The breadth of this language is difficult to overstate. If attached to a policy and applicable to a claim, the exclusion could reach claims involving AI-generated communications, AI-assisted governance decisions, or allegations that the insured failed to adopt adequate AI policies, procedures, or training. The endorsement should therefore serve as a wake-up call for D&O policyholders, including community associations. Associations should ask their insurance professionals whether their current or proposed policies contain an AI exclusion and, if so, how broadly it applies.
AI presents a genuine risk-management dilemma for community associations. Volunteer board members balance association responsibilities with jobs, families, and other obligations. AI can save time by helping organize information, summarize materials, and prepare preliminary drafts. But convenience does not eliminate risk. Generative AI can produce inaccurate, incomplete, biased, or fabricated content. If a board relies on that content when making a governance decision, the association may later have difficulty demonstrating the reasonable inquiry and informed judgment that should support the decision.
AI also raises an explainability problem. A board member should be able to identify the information considered, explain the reasoning behind a decision, and show that the decision resulted from human judgment. A decision becomes harder to defend when the board cannot explain its basis because it relied on an AI system’s opaque output rather than verified facts, governing documents, and appropriate professional advice.
AI can assist with preliminary work, but it is not a substitute for legal advice. Under Sections 3303(a) and 5303(a) of Pennsylvania’s Uniform Condominium Act and Uniform Planned Community Act, board members may rely in good faith on information, opinions, reports, or statements prepared or presented by counsel and other qualified professionals concerning matters within their expertise. An AI platform is not counsel and does not provide the same statutory basis for good-faith reliance. When a governance decision presents legal risk, the board should consult the association’s attorney rather than rely on an AI-generated answer.
Associations should also consider adopting an AI-use policy before a claim arises. Like a social-media or records-management policy, an AI policy can establish practical guardrails for directors, officers, committee members, management, and other association representatives. At a minimum, the policy should:
- define permitted and prohibited uses of AI for association business;
- prohibit entering privileged, confidential, personal, financial, or security-sensitive information into unapproved AI systems;
- require meaningful human review and independent verification of AI-generated content;
- prohibit reliance on AI as a substitute for legal, engineering, accounting, insurance, or other professional advice;
- require governance decisions to remain the product of informed human judgment; and
- identify approved tools, users, record-retention practices, and training requirements.
No AI policy can guarantee insurance coverage. Nevertheless, adopting and consistently following a thoughtful policy can reduce legal and operational risk and help demonstrate that the association exercised reasonable oversight. Associations should also review their D&O, cyber, crime, general liability, and other potentially responsive policies with experienced insurance professionals to identify AI-related exclusions, sublimits, or coverage gaps.
Becker’s Community Association Law and Insurance Recovery teams can assist associations with reviewing AI-related policy language, evaluating governance risks, and preparing an AI-use policy tailored to the association’s operations, governing documents, and insurance program. A short policy adopted now may prevent a much more expensive coverage dispute later.