Is ChatGPT Work Product Protected? Massachusetts Ruling

The use of ChatGPT and other generative artificial intelligence tools in legal disputes is beginning to test the limits of established privilege and discovery doctrines. A recent Massachusetts Superior Court decision addresses one of the more pressing questions: when materials are created using an AI tool in anticipation of litigation, when, if ever, are those materials protected as attorney work product?

 

In Shealy v. Seaside Investments, Judge Debra A. Squires-Lee concluded that ChatGPT queries and AI-generated materials prepared by a litigant’s romantic partner were not protected by the work product doctrine because neither the partner nor ChatGPT was acting as a protected representative of the litigant, and the materials were not prepared at the direction of counsel.

 

Although Shealy is not an employment case, its reasoning will have consequences for employees and employers involved in workplace disputes as courts confront AI-assisted preparation, attorney involvement, and traditional work product principles.

The dispute involved Earl Wade Shealy Jr. and Seaside Investments over a $2 million promissory note. After Seaside sent Shealy a letter asserting that he was in default, Shealy forwarded the loan documents and correspondence to his romantic partner, Debbi Fields.

 

At Shealy’s request, Fields submitted the materials to ChatGPT and used the program to analyze them and prepare a proposed response letter and other documents. Shealy was already represented by counsel at the time, but his attorneys were not involved in the decision to use ChatGPT or in preparing the resulting materials.

 

After Shealy filed suit, the defendants sought production of the communications between Shealy and Fields, including the ChatGPT-generated documents. Shealy objected, arguing that the materials had been prepared in anticipation of litigation and therefore qualified for work product protection.

 

The court rejected that argument and ordered production of the materials.

Generally, the work product doctrine protects certain materials prepared by or for a party or the party’s representative in anticipation of litigation or for trial. Its purpose is not simply to protect anything created after a dispute has arisen. Rather, the doctrine recognizes the importance of protecting litigation preparation, including the mental impressions and strategic analysis of attorneys and other qualifying representatives. That distinction was central to Shealy.

 

The court examined whether Fields could be considered Shealy’s “representative” for purposes of the doctrine. The plaintiffs did not identify any case in which a romantic partner had been treated as a party’s representative under comparable circumstances. Judge Squires-Lee concluded that recognizing such a relationship would expand work product protection without a meaningful limitation.

 

The court likewise rejected the proposition that ChatGPT itself could constitute the relevant representative. The judge distinguished Shealy from Morgan v. V2X, Inc., a 2026 federal case involving a pro se plaintiff who used AI to assist with litigation preparation.

 

In Shealy, the plaintiff had sophisticated legal counsel, but counsel had no role in directing the use of ChatGPT. The court therefore concluded that the AI-generated material did not disclose the mental impressions, conclusions, opinions, or legal theories of an attorney or other protected representative.

 

The court’s characterization of AI was decisive: ChatGPT was treated as a tool, not as a person or legal representative.

Employees involved in workplace disputes often rely on extensive records concerning their employment. Organizing and analyzing that material can become an important part of preparing for a potential claim.

 

AI tools provide one means of doing so, but Shealy demonstrates that the use of an AI program does not, by itself, determine whether resulting materials receive work product protection.

 

For employment lawyers, the more important question may revolve around who directed the work and in what capacity it was performed.

 

There is a meaningful distinction between an employee independently using an AI program to develop an analysis or proposed legal response and materials generated as part of litigation preparation undertaken at the direction of counsel. Shealy does not establish a categorical rule governing every such circumstance, but it indicates that courts may examine the relationship among the litigant, the person preparing the material, the AI system, and the litigant’s attorney.

 

That potential is ripe to be tested in employment litigation, where employees sometimes begin documenting or analyzing a workplace dispute before they have decided whether to retain counsel.

The court’s treatment of Morgan v. V2X, Inc. holds clues to an important distinction between represented and pro se litigants.

 

In Morgan, a federal court in Colorado considered whether materials prepared by a pro se plaintiff using AI could qualify for work product protection. The plaintiff’s lack of counsel was significant because the AI-assisted work was part of the plaintiff’s own effort to develop and present the case.

 

By contrast, in United States v. Heppner, a federal court in the Southern District of New York held that AI-generated materials independently created by a represented defendant did not merit work product protection because they were not prepared at the direction of counsel and did not reflect counsel’s litigation strategy.

 

Shealy falls closer to Heppner. The Massachusetts court emphasized that Shealy was represented by counsel and that his attorneys had no involvement in his decision to have Fields use ChatGPT.

 

For employment disputes, the distinction matters because an employee’s use of AI may occur at several different points in the life of a dispute: before an attorney is retained, during an attorney-client relationship, or while the employee is representing themselves. The legal analysis will likely be different in each circumstance.

Shealy is important not because it establishes that AI-generated materials can never receive work product protection, but because it identifies one particular circumstance in which the doctrine did not apply.

 

Traditional work product principles contemplate preparation by or for a party or a qualifying representative. Generative AI does not fit easily within those categories. An AI system can analyze documents,  summarize evidence, and generate written material, but it is not itself an attorney or human representative.

 

The decision also leaves unresolved whether the use of an open AI platform could create a separate problem concerning confidentiality and waiver of work product protection. The court did not reach that issue because it concluded that the materials were not protected on more fundamental grounds.

 

Those unanswered questions are likely to matter as courts encounter different forms of AI-assisted litigation preparation. There is still considerable room for courts to explore the circumstances under which the AI tool was used, the nature of the material submitted, the identity and role of the person using it, and the extent of counsel’s involvement.

Employment litigation presents particularly fertile ground for these issues because disputes often involve large volumes of documents and detailed factual records. Employees and employers may use AI tools to organize information, review communications, or prepare written material before those materials become part of a discovery dispute, such as after a termination without notice.

 

Shealy suggests that courts are likely to examine the underlying circumstances rather than simply classifying a document according to the technology used to create it. The source of the work, the role of the individual preparing it, and the involvement of counsel may all be relevant to the work product analysis.

 

For employees contemplating litigation, the decision is instructive when it comes to why, how, and when the line between personal preparation and counsel-directed litigation preparation can be decisive. An experienced employment attorney can help assess documents and communications and establish an appropriate framework for preparing a potential claim.

 

Shealy v. Seaside Investments is an early Massachusetts decision addressing a question that will likely arise with increasing frequency in employment and other civil litigation: whether, and under what circumstances, AI-assisted preparation falls within established work product protections. From an employment law perspective, the decision sounds a note of caution surrounding employment disputes, particularly during the early stages when employees may be gathering materials without the support of experienced counsel.

 

If you are in the process of gathering materials for a workplace dispute, schedule a confidential consultation by calling (617) 936-1580 or contacting the attorneys at the Maura Greene Law Group.