Use of Artificial Intelligence in the Community Association Context: Know when to Use A.I. in Conducting your Fiduciary Duties, and When to Leave it to the Experts

Use of Artificial Intelligence in the Community Association Context: 

Know When to Use A.I. in Conducting your Fiduciary Duties, and When to Leave it to the Experts.

Artificial intelligence (“A.I.”), as the recent dramatic increase in its everyday use has shown, can be extraordinarily helpful in many day-to-day tasks if used properly. Models like Chatgpt, Claude, and Gemini do many things, such as, preparing draft communications to the community on difficult topics (like special assessments and rule reminders). A.I. is, at first glance, a very efficient tool to help directors in the governance of their associations. Usually inexpensive or even free, a board can be tempted to use A.I. to carry out tasks typically delegated to a third-party (the managing agent, Association counsel, or another hired expert) and potentially limit those costs. But using A.I. does come with a cost, and the price may be much steeper than those originally avoided by the initial use of A.I.

On the surface, A.I. can create documents that appear to be tailored to an association’s governing documents and which incorporate the applicable common interest community law, but upon closer inspection, A.I. typically fails to account for many necessary considerations and creates something that either exceeds the actual authority of the association and is, therefore, unenforceable, or which fails to capture nuances that allow a given association to take action for which other associations may not have the authority. If a board decides to use A.I. in the course of business, it may find that consulting its human experts on the A.I.’s work reveals significant shortcomings, and it may end up costing the association more money in the long run to have professionals review and correct the A.I.’s work than it would if the board had consulted with its experts initially and had them do the work.

Although many A.I. products have seemingly “unlimited access” to information, those products may not have the capacity to discern when certain information is applicable. For example, in order to draft a policy resolution, the board must have the authority to adopt rules and regulations pertaining to the topic of the policy. Some topics may be regulated by statute and therefore may limit rules on such topics, or may provide additional authority not found in the governing documents. In addition, case law may provide the board with additional or restricted authority as to a given topic. A.I. is only as good as the information that it can access and while the information available to A.I. seems unlimited and is ever-expanding, it typically does not have all of the information that is relevant to an association’s inquiry. Even if A.I. may be able to find these nuanced sources, again it may not be able to incorporate them into a policy appropriately.  In the end, a board may draft a policy using A.I. that exceeds its authority or which omits key authorities that the board may otherwise have (and therefore unnecessarily limits their authority).  The resulting A.I. work product may be either unenforceable or weak. To avoid these issues, a board may send the draft policy to its legal counsel for review. In our experience, however, this review can end up costing more than if the board had authorized counsel to draft the policy initially.

Reviewing A.I.-generated materials can end up costing the association more in the long run because of the steps needed to ensure the contents are accurate. First, the initial policy must be reviewed for both accuracy of content (what the regulations are) and the basis of authority cited for such regulations (why the association can regulate this topic in the proposed manner). To ensure such accuracy, the policy must be reviewed closely against the association’s governing documents and applicable law. Often, A.I.-drafted materials are overbroad and make assumptions that do not apply to the association, and therefore must be significantly tailored to reflect the actual authority the association possesses.

Second, the materials often may need to be re-worded to ensure consistency with the association’s other documents. For instance, the A.I. work product may incorrectly use words that are different than the defined terms in the association’s declaration or bylaws.  The A.I.’s formatting may be inconsistent with other documents published by the association and therefore warrant reformatting to ensure consistency and to avoid questions as to a document’s authenticity.

Third, A.I.-produced content can often simply be wrong, frequently misstating or misunderstanding nuances of the law or documents, or sometimes hallucinating law or incorporating information from other associations or association types that do not apply. Parsing out the incorrect information, and correcting it, is frequently more time consuming than using a knowledgeable drafter starting from scratch.

In the context of answering questions rather than drafting documents, A.I. is frequently mistaken. Occasionally, board members or managers will ask an A.I. chatbot a question before sending it to our firm, and they ask “is this correct?” and often the answer is “no.” Fortunately for those who consult with legal counsel on the A.I. response, we are able to provide them with the correct information before they take action based on that incorrect information provided by A.I.

Finally, the use of A.I. presents potential increased exposure for individual directors that choose to rely on it for information and guidance. A.I. may appear to be an expert in many things, but the law does not provide an exception or insulation from liability to a Board who uses such and gets bad advice. Any director that relies on A.I. in the same manner that the director might rely on professional experts does so at their own risk and peril. Virginia law offers protections for directors who rely on the advice of experts, but does not offer the same protection for directors who rely on the advice of A.I. Moreover, A.I. has no duties to the user or to the association in the same way that professionals do. It also does not have errors & omissions insurance like many professionals are required to carry. In this way, relying on professionals rather than A.I. serves to protect directors from liability and provides additional protections to the association in case that reliance is misplaced.  Also keep in mind that when you feed information into many A.I. tools, those A.I. tools retain and use that information for other purposes.  In February 2026, a federal judge in New York ruled that by inputting information from the attorney into a public A.I. platform that such action acts as a waiver of attorney-client confidentiality because that A.I. platform is a third party, not legal counsel, and there is no expectation of privacy.

A.I. is a useful tool but due to its limitations, it may not always be an appropriate choice to handle association matters. To ensure the association is acting within its own authority, and publishing documents and taking positions on issues that reflect the same, directors should consider whether the costs that may be incurred due to using A.I. in the long run outweigh the benefits that may result from its immediate use for a quick, easy “answer.” Guiding this analysis should be the director’s commitment to upholding their fiduciary obligations. At the end of the day, A.I. may seem like the board’s best friend, but it can also be the wallet’s worst enemy and get boards into more trouble without proper oversight.