Massachusetts court limits work product protection for AI output

A Massachusetts Business Litigation Session judge ruled in Shealy v. Seaside Investments, LLC that AI-generated materials are not protected by the work product doctrine unless they are created for counsel or at counsel’s direction. The decision treats AI output under traditional privilege and work product rules, rather than creating a separate standard for AI tools.

The dispute arose from a commercial investment relationship involving a purchase agreement and promissory note executed in January 2025. After Seaside sent correspondence in May 2025 about a potential default, Earl Wade Shealy, Jr. forwarded documents to his romantic partner, Debbi Fields, who uploaded them to ChatGPT with prompts and returned the generated output. Neither Shealy nor Fields consulted an attorney during that process, although Shealy was represented by counsel in the transaction.

The court found that Fields and ChatGPT were not representatives covered by the work product doctrine, and that the AI output did not reflect the mental impressions, conclusions, opinions, or legal theories of an attorney or qualified representative. The ruling distinguished cases protecting AI prompts and outputs used by pro se litigants, emphasizing that Shealy acted independently while represented by counsel.

The decision signals that parties cannot route litigation-related analysis through non-attorneys or AI tools and assume the results will be shielded from discovery. Counsel-directed AI use may be treated differently, but independent use can leave both prompts and outputs exposed.