In a decision of first impression in Massachusetts, a Superior Court judge has ruled that AI-generated output does not qualify for work product protection unless directed by counsel.
The case, Shealy v. Seaside Investments, LLC, centres on a commercial investment relationship in which parties executed a purchase agreement and promissory note in January 2025.
The dispute escalated after Seaside sent correspondence in May 2025 concerning a potential default, prompting Earl Wade Shealy, Jr. to seek an unusual form of analysis.
Rather than consulting his attorney, Shealy sent Seaside’s correspondence and related transaction documents to his romantic partner, Debbi Fields, to have ChatGPT analyse the materials and draft a response.
Fields uploaded the documents to ChatGPT with instructions and prompts, then returned the AI-generated output to Shealy, with no attorney involved at any stage of the process.
When litigation commenced, Shealy produced redacted correspondence between himself and Fields, claiming the AI-generated materials deserved work product protection shielding an attorney’s mental impressions and legal theories from discovery.
The court rejected that argument on multiple grounds, finding that neither Fields nor ChatGPT qualified as a “representative” under the work product doctrine as applied in Massachusetts.
The court stated directly: “AI is a tool not a person. Nothing about what ChatGPT produces discloses the ‘mental impressions, conclusions, opinions, or legal theories of an attorney or other representative of a party concerning the litigation.'”
The ruling distinguished cases from federal courts in Colorado and Michigan, including Morgan v. V2X, Inc. and Warner v. Gilbarco, Inc., where AI-generated output was protected for pro se litigants acting without legal representation.
Those courts, including the Morgan court, extended protection partly to level the playing field between represented and unrepresented parties, a rationale the Massachusetts court found inapplicable to Shealy’s situation.
Because Shealy was represented by counsel at the time he used ChatGPT through a third party, the court found he had acted entirely on his own initiative without consulting that counsel.
The court also drew on the February 2026 decision in United States v. Heppner, where a represented criminal defendant’s independent AI use similarly fell outside privilege and work product protection.
Massachusetts does not recognise romantic partners as representatives for work product purposes, meaning exchanges between Shealy and Fields were fully subject to disclosure to the opposing party.
The decision signals a firm judicial position that AI tools, however powerful, do not inherit the protections traditionally afforded to attorneys or their qualified representatives in litigation settings.
For businesses and individuals anticipating litigation, the ruling delivers an unambiguous message: AI-assisted analytical work must be directed by counsel to have any realistic chance of remaining shielded from discovery.