A self-check for the managing partner or administrator who has to answer the question — "are we using AI responsibly?" — in plain English, without becoming a technologist to do it.
In United States v. Heppner (S.D.N.Y., Judge Rakoff, decided Feb 11, 2026), a federal court held that one defendant's exchanges with a consumer AI platform were not protected by attorney-client privilege or work-product. The holding rested on three specific grounds: the AI is not an attorney; the provider's privacy terms defeated confidentiality; and the use was not at counsel's direction.
The risk it spotlights is narrow and specific: unsupervised, counsel-undirected, consumer-tool use — not "all AI waives privilege." The court noted that AI use directed by counsel may remain protected (the Kovel doctrine). The point isn't to avoid AI — it's to use it in a way you can document.
The ABA's first generative-AI ethics opinion sets affirmative duties for lawyers using these tools: competence (Rule 1.1), confidentiality (Rule 1.6 — the opinion is clear that boilerplate engagement-letter consent is not adequate), client communication (1.4), candor (3.1/3.3), supervision of lawyers and staff (5.1/5.3), and reasonable fees.
What this self-check does: eleven plain-English questions map your firm's current posture against those duties and the Heppner fact pattern, then return a readiness band and an honest picture of what firms at your level are usually documenting next.
Answer for your firm as it actually operates today, not as you intend it to. "Unsure" is a valid and useful answer — an undocumented practice is its own kind of gap.
This band reflects documentation and process maturity only. It is not a legal conclusion, a compliance certification, or a risk rating from your carrier — only your counsel and bar can speak to those.
This decides whether a live discovery call or a self-serve guide is the better fit for you. (Still scored only in your browser.)