Texas court protects generative AI review prompts

A federal magistrate judge in Austin declined to order securities defendants to turn over the prompts they used with a generative AI review tool, treating attorney-crafted instructions as protected work product. The outcome gives lawyers an early marker in disputes over whether generative AI inputs used in legal review must be disclosed during discovery.

The ruling’s practical value is limited by its form. Rather than a written opinion laying out the court’s reasoning, the decision appears only as a docket entry and in a brief hearing transcript. That leaves practitioners with a notable result but little guidance on how courts may analyze prompt disclosure, privilege, and work product protections in future AI-related discovery fights.