GenAI + Privilege: Public Tools Can Expose Legal Work

Belief: using a public GenAI tool for legal work is like using a search engine, so privilege stays intact.

Privilege depends on confidentiality. If a tool’s terms allow retention or disclosure, your prompts and outputs can become discoverable, and the privilege story gets harder.

The break point is casual experimentation. Someone drafts an internal analysis, summarizes sensitive notes, or prepares talking points using a consumer AI tool. Later, the company tries to treat the output as privileged because counsel became involved afterward.

The proof pressure point is the platform and the workflow: whether the tool was controlled, whether use was directed within a legal process, and whether confidentiality was realistically preserved.

The practical fix is platform discipline. For legal work, use enterprise tools with contractual protections and treat AI use as a records-creating workflow that needs rules, not improvisation.

This post shares general information based on common patterns I see in California workplaces. It is not legal advice, does not create an attorney-client relationship, and outcomes depend on specific facts — no lawyer can guarantee a result. Past results do not guarantee or predict future outcomes. AI may have been used to create this post. All content reviewed by a CA attorney before publication. This post may be attorney advertising.

Michael Trust Law, APC, 703 Pier Avenue, Ste. B367, Hermosa Beach, CA 90254: michaeltrustlaw.com