Legal
How Genflare uses AI in its own writing, and what that means for you as a client or reader
Genflare uses AI at every stage of our writing: to research, to think against, to structure, to draft, to challenge a draft, and to check facts. The degree of assistance varies from piece to piece and from sentence to sentence. What does not vary is the standard: a named person at Genflare reads everything before it leaves, approves it, and stands behind it as its author. If no one will sign it, it doesn’t ship.
We treat authorship as an act of judgment, not of typing. The author of a Genflare document is the person who set its intent, made its governing choices, verified its claims, and approved its final form. That person is named on the work.
Every deliverable carries a named author. Not “the Genflare team” and never “the AI.” A person you can call.
Every checkable claim traces to a source. Figures, quotations, regulatory references, and statements about your business are verified against a source before they reach you. The model is never the source. When we cannot verify something, we say so rather than smooth it over.
Certain content is decided by a person, every time. AI may help us structure and clarify it; it never determines what is said. This includes any assessment of your organization, your people, or your competitors; recommendations you will act on; contract, pricing, and engagement terms; and anything to be attributed to a named individual, ours or yours.
It says what we mean. We do not send polished text no one at Genflare meant. If a sentence is in a Genflare document, someone here chose it.
A standing policy is a better form of transparency than a footnote on every message, so this page is our primary disclosure. Where a specific piece of writing warrants it — client deliverables, published work — it carries a short declaration in the author’s voice. The house standard reads:
I use AI for research, drafting, and revision. Whatever the degree of assistance, everything I send is directed, reviewed, and approved by me — it says what I mean to say, and I am its author.
Routine correspondence is covered by this policy and carries no separate line. If you ever want to know how a particular piece of our writing was made, ask the author; we will tell you.
We use AI tools under business terms that do not train on the material we provide. Client confidential material enters those tools only as far as the engagement requires, and never into consumer-grade or personal accounts. Where an engagement’s terms are stricter than this policy, the engagement’s terms govern.
We do not run AI-detection software on anyone’s writing — not clients’, not partners’, not our own people’s. Detectors are unreliable and their errors fall on the wrong people. We judge writing by its accountability and its quality, never by guessing at its provenance.
When we write or help design patient-facing communication for a client, we do not use AI to personalize it to a patient’s psychology in order to move their behavior. Patient communication is personalized to what the reader needs to understand — their coverage, their language, their moment — never to what might persuade them. This is a fundamental rule of our practice, and engagements that touch patient materials invoke it explicitly.
We do not let volume outrun review. If we cannot read it, we do not send it.
This is a plain-language statement of practice, not a regulatory designation. Good Writing Practice is Genflare’s internal name for this standard; it implies no regulatory status. Our records of how a given document was made are, today, the author’s attestation under this policy; as our tooling matures we intend to capture more of that record automatically, and we will update this policy only as our records can back it.
Questions about this policy go to its owner at greg@genflare.ai.
Genflare, Inc. · San Francisco · Published alongside our Privacy Policy.