Bottom line: Yes, most agencies should disclose AI use to clients, and do it upfront. Be clear-eyed about the cost: research from the University of Arizona's Eller College of Management found that people who disclose AI use are trusted less, and softer wording did not remove that penalty. But the same research found that being exposed by someone else is worse than disclosing yourself. Add contracts, confidentiality and a growing set of rules, and upfront disclosure is the lower-risk choice. The work is choosing the right level of detail for each client and putting it in writing.
Most agency owners ask "should I disclose?" The more useful questions are how to disclose, to which clients, in what level of detail, and what the contract should say. This post answers all four: what the research actually found, a 3-tier disclosure decision, the rules that apply as of October 2026, and a contract clause you can adapt.
Quick-scan summary:
- Default answer: yes, disclose, early. Clients who find out later, or hear it from someone else, tend to react worse than clients who were told at the start.
- The research: Eller College researchers ran 13 experiments with more than 5,000 participants. Disclosing AI use lowered trust in every scenario, including about 20% lower trust in graphic designers who disclosed it. Framings such as "a human reviewed the AI's output" did not remove the drop. Third-party exposure lowered trust even more.
- 3-tier disclosure decision: Tier 1 (full transparency, the default for most agency work), Tier 2 (selective, per-deliverable disclosure for regulated industries and sensitive content), Tier 3 (quiet integration for internal tooling that never touches the deliverable).
- Rules to know: the EU AI Act's Article 50 transparency obligations have applied since August 2, 2026; the FTC's rule banning fake reviews and testimonials, including AI-generated ones, took effect in October 2024; California's AI Transparency Act (SB 942) applies to large generative AI providers, not directly to agencies.
- Contract language: add an "AI tools use" clause covering which tools you use, what data goes into them, who reviews output, and who owns the IP. Sample clause below.
Why this question is different in 2026
Two things changed between 2023 and 2026:
- AI output got hard to spot. AI detectors are unreliable, so clients often cannot tell from the output alone. Disclosure is now a choice you make, not something the work forces on you.
- Rules caught up in specific areas. The EU AI Act's transparency obligations now apply, the FTC has a rule against AI-generated fake reviews, and regulators in finance and law have said their existing rules apply to AI-assisted work.
The era of quiet AI use without any client conversation is ending. Agencies that set clear disclosure norms early have fewer awkward conversations later.
What the research actually says
The most cited study is from the University of Arizona's Eller College of Management. Oliver Schilke and Martin Reimann's paper, "The Transparency Dilemma: How AI Disclosure Erodes Trust," was published in Organizational Behavior and Human Decision Processes in 2025. Eller's summary of the research reports:
- 13 experiments, more than 5,000 participants. In each, people who disclosed AI use were trusted less than people who did not.
- The drop was material. Students trusted a professor 16% less after learning AI was used for grading. Investors trusted firms 18% less when ads disclosed AI use. Clients trusted graphic designers 20% less after AI disclosure.
- Softer wording did not fix it. Framings such as "AI was only used for proofreading" or "a human reviewed the AI's output" still produced a trust decline.
- Being exposed is worse. Trust fell further when a third party revealed the AI use than when the person disclosed it themselves.
What this means for agencies: you cannot word your way out of the disclosure penalty, so do not promise your team a magic phrase. What you can control is timing (upfront, before anyone else raises it), specificity (what AI does and does not do in your process), and accountability (a named person owns every deliverable). Then let the quality of the work carry the relationship.
The 3-tier disclosure decision
Every piece of agency work falls into one of three disclosure tiers. Pick the tier based on the work and the client.
| Tier | When to use | What to disclose | Where to disclose |
|---|---|---|---|
| 1. Full transparency | Default for most agency work | "We use AI tools for research, drafts, and analysis. Final work is reviewed and approved by our team." | Onboarding, master agreement, sales conversations |
| 2. Selective disclosure | Regulated industries, sensitive content, premium creative work | Specific AI uses by deliverable type, with explicit "AI was not used" flags on creative judgment work | Per deliverable, in proposals, in regulated-content audit trails |
| 3. Quiet integration | Internal tooling that does not feed the deliverable | Nothing client-facing, but documented internally | Internal SOPs and tooling docs |
The decision flow:
- Is this regulated content (medical, legal, financial, pharmaceutical, government)? → Tier 2, always.
- Is the client paying a premium for human creative judgment (brand strategy, complex creative, senior advisory)? → Tier 2.
- Is AI only in your internal toolchain and not in the deliverable (for example, AI summarizing internal research that a person turns into a written brief)? → Tier 3.
- Everything else → Tier 1.
For most agencies, Tier 1 covers the bulk of client work.
Tier 1: Full transparency (default)
What you sell: marketing execution where AI speeds up production but people direct, review, and own the output.
How to describe it: say what AI does in your process and who is accountable. Vague statements invite worst-case assumptions. Avoid:
- "We use AI to write your content."
- "AI generates our reports."
Use instead:
- "Our team uses AI for first drafts and research aggregation. A senior strategist reviews, edits, and approves every piece before it reaches you."
- "We use AI to cut report assembly time by 60% so our analysts spend their time on interpretation and recommendations." (Swap in the figure you have actually measured.)
The research says even good wording carries some trust cost. Specific wording at least gives the client accurate expectations, which matters more over a long engagement than a first impression.
Where to disclose:
- Onboarding doc: a one-paragraph "How we use AI" section in the welcome materials new clients receive.
- Master agreement: a contract clause (template below) covering AI use, data handling, and IP.
- Discovery calls: answer plainly when a prospect asks about your process or quality control. Do not lead with it (you are selling outcomes, not AI), but do not dodge it.
The goal is for the buyer to see a responsible operator with good tools, not an AI shop with no people behind it. This connects to the broader agency positioning question.
Tier 2: Selective disclosure
For regulated and high-trust work, you need per-deliverable disclosure. The client may need the record for its own compliance or governance.
When this applies:
- Healthcare and pharma: promotional materials go through the same regulatory and medical-legal review however they were drafted, and reviewers will want to know where AI was involved.
- Financial services: FINRA's Regulatory Notice 24-09 (June 2024) says the content standards of its communications rule apply whether a communication is generated by a person or a technology tool.
- Legal services: the American Bar Association's Formal Opinion 512 (July 2024) says lawyers using generative AI must protect client confidentiality and may need a client's informed consent before entering confidential information into some AI tools. Law-firm clients will often pass these expectations on to agencies.
- Government and public sector: procurement rules and AI policies vary by agency and jurisdiction.
- Brand strategy and senior creative for clients who explicitly pay for human judgment.
What you disclose:
- Which AI tools (and versions) were used
- What data went into them (no client data, anonymized data, or client data under an approved tool)
- What stage of the work AI touched (research, draft, final, none)
- Who reviewed and approved
- Which parts were human-only
Where you disclose: a deliverable-level record, similar to a creative production log, that the client can review on request.
Tier 3: Quiet integration
Some AI use does not need client-facing disclosure because it is internal tooling that never appears in the deliverable.
Examples:
- AI summarizing internal research when the deliverable is a human-written brief
- AI for SOP documentation and team training
- AI for prospect research and lead qualification (internal sales)
- AI for proposal drafting where senior staff then rewrite the proposal
The principle: if a reasonable client would not consider the AI use material to the work they bought, internal use is fine without per-deliverable disclosure. Document it internally so you can answer if asked.
Caveat: Tier 3 is shrinking as client expectations and contract terms tighten. In 2026 the conservative move is to push borderline cases into Tier 1.
What the rules require (as of October 2026)
This is not legal advice. Check with counsel for your jurisdictions and clients.
EU AI Act: The Act entered into force in August 2024, and its Article 50 transparency obligations have applied since August 2, 2026. The parts most relevant to agencies apply to "deployers" (organizations using AI): content that is a deep fake (realistic AI-generated or manipulated images, audio or video of people, places or events) must be disclosed as such. AI-generated text published to inform the public on matters of public interest must also be disclosed, unless it has gone through human review and someone holds editorial responsibility. Separate marking duties fall on the providers of generative AI systems. If you produce content for EU audiences, map your work against these categories.
FTC (US): There is no general federal rule requiring businesses to disclose AI use. The FTC Act's ban on deceptive practices still applies to AI-assisted marketing, and the FTC's final rule on fake reviews and testimonials, announced in August 2024 and effective October 21, 2024, prohibits fake reviews and testimonials, including AI-generated ones attributed to people who do not exist. In September 2024 the FTC also announced "Operation AI Comply," a set of enforcement actions against deceptive uses of and claims about AI.
California SB 942 (AI Transparency Act): This law applies to "covered providers," meaning companies that produce generative AI systems with more than 1,000,000 monthly users that are publicly accessible in California. It requires them to offer a free AI-detection tool and to embed disclosures in AI-generated image, video and audio content. AB 853 moved its operative date to August 2, 2026. It does not directly require agencies to label their work, but the tools your team uses may now mark their output.
Industry-specific: FINRA and the ABA guidance above apply existing rules to AI-assisted work. Expect regulated clients to write AI terms into their vendor contracts.
Practical implication: for most agencies, the strongest obligations come from client contracts (confidentiality, data use, IP warranties) and from specific content types (deep fakes, fake testimonials, regulated communications), not from a blanket disclosure law. That is why the contract clause below matters.
Sample contract clause (adapt and use)
Add this to your master services agreement and have your lawyer review it for your jurisdiction.
AI Tools Use Clause
Service Provider may use generative AI tools and machine learning systems as part of its production workflow, including for research aggregation, first-draft content generation, analysis, and process automation. Service Provider warrants that:
- All deliverables undergo human review and approval by qualified Service Provider personnel before delivery to Client.
- No Client confidential information will be submitted to public AI tools without prior written approval. Service Provider maintains an approved-tools list available to Client on request.
- Service Provider retains responsibility for all deliverables regardless of which tools were used in their production.
- Where required by applicable law or industry regulation (including but not limited to the EU AI Act, FTC rules, and state-specific AI laws), Service Provider will identify AI-generated content per the applicable disclosure requirements.
- Intellectual property in deliverables transfers to Client per the IP terms of this Agreement, regardless of whether AI tools were used in production.
Client and Service Provider agree that AI tool use does not by itself constitute a substitution of services and does not modify the obligations or deliverables of either party under this Agreement.
The clause does five things: confirms AI use, sets the data-handling boundary, keeps accountability with the agency, commits to legally required disclosure, and clarifies IP ownership. One caution on point 5: in some jurisdictions, including the US, purely AI-generated material may not be protected by copyright, so ask counsel how to word IP warranties for AI-assisted work.
The "just don't disclose" trap
A common 2026 mistake: quietly using AI without telling clients because "they'll never know."
Why this fails:
- They may find out anyway. Through a junior team member's offhand comment, a turnaround time that does not add up, a contract audit, or a tool that now marks its output. The Eller research found trust falls further when a third party exposes AI use than when you disclose it yourself.
- The contract is exposed. If your master agreement does not address AI use and the client discovers it, questions about IP and confidentiality (what data went into which tool) become a real problem.
- Your team is put in a bad spot. A junior team member asked "how did you produce this so fast?" has to either mislead the client or contradict you. Both hurt trust.
Disclosing well costs a paragraph in onboarding, a contract clause, and some discovery-call conversations. Being found out can cost a client and start a contract dispute. The risk is lopsided, and it favors disclosure.
Not for you
This playbook is not for you if:
- You run an AI-first agency where AI is the core of what you sell (the AI implementation lane in the 4-lane positioning test). Disclosure is your positioning, not a policy question.
- You serve only government or defense contracts. Those requirements are jurisdiction-specific and stricter than this post covers. Get specialist counsel.
- You operate mainly outside the US and EU. Local rules vary; you need local counsel.
It is for you if you run a services agency in the US, EU, UK, or a comparable jurisdiction and want a defensible default position on AI disclosure.
FAQ
Do companies have to disclose if they use AI?
Sometimes. In the EU, Article 50 of the AI Act (applying since August 2, 2026) requires disclosure of deep fakes and of AI-generated text published to inform the public on matters of public interest, unless that text has had human review and editorial responsibility. In the US there is no general federal disclosure rule, but the FTC prohibits deceptive practices, including AI-generated fake reviews and testimonials. Regulated industries such as financial services apply their existing communication rules to AI-assisted content. Outside those cases, disclosure is mostly a contract and trust decision.
Do you have to tell clients you are using AI?
In the US, no federal law requires general agency-to-client AI disclosure. Your contracts may, though: confidentiality, data-use and quality-warranty clauses can all be triggered by AI use, and regulated clients may add AI terms to vendor agreements. The practical answer is to tell clients upfront. Research from the Eller College of Management found that trust drops further when someone else exposes AI use than when you disclose it yourself.
Should I disclose my use of gen AI to clients?
Yes for most client work, with specifics. Eller College research found that disclosure lowers trust somewhat and that softer wording does not remove the effect, so the case for disclosing rests on avoiding the bigger loss from being found out, meeting contract and legal obligations, and setting accurate expectations. Use full transparency for most work (Tier 1), per-deliverable disclosure for regulated industries (Tier 2), and internal documentation only for AI that never touches the deliverable (Tier 3).
Can a company tell if you used AI?
Not reliably from detection software, which produces false positives and misses. Clients more often notice from volume (a 20-page report delivered in 3 hours), from patterns (generic phrasing, repetitive structure), or because someone on the agency side mentions it. Some AI tools also now embed provenance data in images, video and audio.
What should be in an AI-use clause in agency contracts?
Five elements: (1) acknowledgment that AI tools may be used in the workflow, (2) a data-handling boundary (no confidential client data in public AI tools without approval), (3) human accountability for all deliverables, (4) compliance with applicable AI disclosure laws, and (5) IP terms that apply regardless of AI use. The sample clause earlier in this post covers all five.
Do clients trust agencies less when they know AI was used?
Often, yes, at least initially. In the Eller College experiments, clients trusted graphic designers about 20% less after the designers disclosed AI use, and softer framings did not remove the drop. Trust fell further when a third party revealed the AI use. That is why the advice is to disclose early and specifically, and let consistent work quality rebuild confidence, rather than to hide it.
How should small agencies handle AI disclosure?
The same way large agencies should: a clause in the master agreement, a one-paragraph "How we use AI" section in onboarding materials, and a clear answer in discovery calls. Small agencies often skip the contract step because they use lightweight agreements, which leaves them the most exposed. A short addendum to your existing agreement closes the gap.
What to do next
To adopt a defensible AI disclosure stance:
- Add the AI tools use clause (above) to your master agreement, with counsel review.
- Write a 100-200 word "How we use AI" paragraph for client onboarding.
- Train your team to explain AI use plainly: what it does, what it does not do, who reviews the work and who is accountable.
- Audit your client roster: which clients are in regulated industries that need Tier 2?
- Read the broader context in Will AI replace marketing agencies? and How to position your agency in the AI era.
Agencies that settle this now will have a ready, consistent answer when a client asks.
Related reading:
