What the Client Deserves to Know: A Tiered AI Disclosure Standard for Independent US Studios

Independent design studios in the United States have entered a period in which the provenance of work is part of the work. Clients want to know whether a campaign concept, a website interface, or a brand system was shaped by generative AI, and they are asking earlier and more specifically than most studios anticipated. The answer cannot be assembled from memory at the closeout meeting. It needs a habit.

The cost of improvising an answer

A client asks, midway through a brand refresh, whether the copy in a campaign concept was generated by AI. The designer answers from memory. A second client asks at closeout whether any images were synthetically produced. The answer differs in tone, detail, and confidence. Neither answer is dishonest, but neither is a policy. The cost is not only legal exposure; it is the slow erosion of a studio’s reputation for steadiness. Independent practices win work on judgment, and judgment includes knowing what you did and being able to say it plainly.

A disclosure habit solves a narrower problem than a full AI governance program. It gives the studio a small set of answers it can stand behind, applies them consistently, and documents the basis for each answer. The goal is not to narrate every keystroke. The goal is to make the studio’s use of AI legible to clients who have legitimate reasons to ask.

Three tiers, plainly defined

A workable standard begins with categories that a client can understand without a glossary. FNAD recommends that studios define three disclosure tiers and apply the highest tier that touched the deliverables.

Tier 1 — No generative AI in deliverables. The studio may use conventional software, spell-check, and other tools that do not generate client-facing content. No generative AI system was used to produce or materially revise the delivered work.

Tier 2 — AI-assisted process, human-authored output. Generative AI was used for internal exploration, summarization, research triage, or editing suggestions, but the delivered content was authored, selected, and revised by the studio. The tool did not produce final client-facing text, imagery, or code.

Tier 3 — AI-influenced or AI-generated output. Generative AI produced or materially shaped client-facing content, including draft copy, image elements, code, or design variations that appear in the deliverables. A human designer reviewed and edited the output, and the studio can identify the stage and category of tool involved.

The tiers are deliberately broad. A studio that tries to list every product name will find the list stale within months. Categories such as text generation, image generation, and code assistance travel better. The tier should be recorded at project close and updated if the workflow changes.

Contract language that travels

Disclosure works best when it is expected before the first invoice. A short rider or master services clause can set the terms without turning the studio into a compliance department. The following language is a starting point, not legal advice; studios should have counsel review it for their jurisdiction and client mix.

Studio will disclose in writing, at project close or earlier upon client request, the highest AI disclosure tier applicable to the deliverables. The disclosure will identify the category of tool and the stage of work affected, without requiring disclosure of proprietary prompts, internal drafts, or workflow details.

Client may request a Tier 1 workflow for designated deliverables. Such requests will be scoped, scheduled, and priced separately, and may require changes to timeline or scope.

Studio will not input client confidential materials into generative AI tools unless the client has approved the tool category in writing.

Those three sentences do most of the work. They establish that disclosure is routine, that a no-AI request is possible but not free, and that confidentiality is a separate gate. The clause should appear in the master agreement and be referenced in project proposals so that clients see it before they sign.

Documentation without bureaucracy

A disclosure policy fails when it depends on memory. It succeeds when the record is small enough to keep. For most independent studios, a single project log entry is sufficient. Record the date, the category of tool, the stage of work, whether client confidential material was involved, who reviewed the output, and the final disclosure tier. Store the entry with the project file. If the studio uses a project management system, add a required field to the closeout checklist.

The log should not become a diary. A sentence per use is enough. The purpose is to support a consistent answer if a client asks six months later, or if a new team member inherits the account. Documentation also helps the studio notice patterns: if Tier 2 use is creeping into every project, the studio may need to revisit its default assumptions about speed and scope.

A one-page internal policy

The most useful policy is the one a busy principal will actually read. FNAD suggests a one-page document with seven numbered items:

1. Default tier for new work, and who may approve a change.
2. The intake question that asks clients whether they have AI restrictions.
3. The written disclosure requirement at project close.
4. The prohibition on confidential inputs without written approval.
5. The human review requirement for any Tier 3 output.
6. The contract rider or clause reference.
7. An annual review date, with a named owner.

The policy should be dated and signed by the principal. It does not need to be published on the website, though some studios will choose to share it. Its primary audience is internal: it tells a freelancer, a junior designer, or a new project manager what the studio means when it says it uses AI responsibly.

The trust ledger

Disclosure is not a concession to suspicion. It is a form of professional bookkeeping. Clients who ask about AI are usually asking a larger question: can I trust this studio to tell me the truth about my work? A tiered standard answers that question before it becomes a negotiation. It also gives the studio a defensible position when a client requests a no-AI workflow, because the request can be scoped rather than resisted.

For independent US designers, the competitive advantage is not secrecy. It is clarity. A studio that can say, in plain language, what it used and what it did not, is a studio that has turned a new source of anxiety into a routine part of professional practice. FNAD’s role is to keep that routine practical: small enough to sustain, specific enough to mean something, and flexible enough to survive the next tool.