AI Disclosure: Should You Tell Customers When AI Wrote It?

The short answer is yes, in most contexts, and mostly for business reasons rather than legal ones. There is no single US law requiring a “this was written by AI” label on general marketing content or blog posts. But 89 percent of small businesses now use AI in some form, according to Capsule CRM and the SBE Council in 2026, which means the disclosure question is not theoretical anymore. Most businesses skip it. A growing share of customers notice when something feels off, even if they cannot say why. Here is what the legal landscape actually looks like, and why the argument for transparency tends to be stronger than the legal minimum suggests.

What the law actually requires right now

The legal picture is less settled than most articles suggest, and that matters before you build a policy around it.

No blanket US mandate covers general content. There is no federal law requiring a disclosure label on AI-drafted blog posts, marketing emails, product descriptions, or website copy. The FTC’s general deception standard applies: if a piece of content would mislead a reasonable consumer about its origin, the FTC expects disclosure. But there is no AI-specific rule for ordinary business writing.

Advertising and sponsored content are a different matter. The FTC’s endorsement guidelines require disclosure when a material connection might not be obvious to consumers. Applied to AI, this means sponsored posts and paid placements that lean heavily on AI-generated text likely require disclosure, especially when a reader might reasonably believe they are reading genuine human experience or opinion.

New York and the EU added specifics in 2026. New York’s disclosure law took effect June 2026 and targets AI-generated advertising content featuring synthetic performers meant to look human. The EU AI Act, which took full effect on August 2, 2026, requires disclosure when AI-generated text touches matters of public interest, including business and financial content reaching EU audiences.

The landscape is fragmented and moving quickly. The safest default is to establish a clear disclosure standard now and adjust as specific rules develop, rather than waiting for a consolidated answer that may be years away.

One figure worth noting: 96 percent of AI Overview citations in search come from sources with strong E-E-A-T signals, according to GEO research in 2026. The legal floor is the minimum. How you build trust determines the ceiling.

The business case for disclosure is stronger than the legal floor

The most compelling argument for disclosure is not that you are required to do it. It is that transparency is a competitive signal.

Customers in 2026 are more AI-aware than most businesses give them credit for. Consumer use of AI tools to research and find businesses jumped from 6 percent in 2025 to 45 percent in 2026, according to Cheers research. People who use AI search daily develop a feel for content that is smooth but oddly imprecise. A disclosure removes that doubt and replaces it with something more useful: a signal that this business is thoughtful about how it uses AI, which usually means it is thoughtful about other things too.

There is a search visibility angle as well. AI search engines, including ChatGPT, Perplexity, and Google’s AI Overviews, now answer questions about businesses and services at scale. Only about 1.2 percent of businesses get recommended by ChatGPT at all, according to SOCi research from 2026. The businesses that do tend to have strong E-E-A-T signals. A transparent disclosure, paired with clear human authorship and visible expertise, contributes to that signal in a way that silence does not.

For businesses building AI automation and content workflows, there is an operational angle too. If AI is involved in content your team produces, having a stated policy rather than improvising case by case is what makes the practice manageable as you scale. The post on why AI team rollouts stall covers this from the adoption side: informal use without a consistent framework tends to create the kind of inconsistency that shows up as a credibility problem over time.

What a good disclosure actually looks like

The disclosure does not need to be prominent or confessional. It needs to be accurate, specific, and positioned near the content it describes.

For blog posts and articles, a single line near the author block or at the foot of the piece is enough. “This article was drafted with AI assistance and reviewed by our team” is clear and honest. Vague alternatives like “AI-enhanced content” or “written with smart technology” do not tell readers anything real and can read as evasive.

For marketing emails, the threshold is lower. Conversational email is not usually subject to the FTC’s ad-specific rules unless it functions as paid placement or sponsored content. That said, if your newsletter reads as your direct voice and AI contributed most of the structure, a brief footer note costs nothing.

For client deliverables, proposals, reports, and strategy documents, the stakes are meaningfully higher. These are purchased as professional judgment. The disclosure is not a footer line; it is a conversation before you deliver. Google’s March 2026 core update targeted scaled AI content without meaningful human review, with affected sites seeing 50 to 80 percent traffic drops, according to Digital Applied. The same principle holds in client relationships: what separates AI-assisted content from AI-generated content is the human review step. The disclosure is partly how you tell clients that step actually happened.

The pattern that consistently backfires is vague language. “AI-assisted,” “smart tools,” “technology-enhanced writing” give readers no real picture. If you are going to disclose, be specific enough to be useful.

Where disclosure matters most (and where it matters less)

Not every use of AI requires the same treatment. The question worth asking is: if a customer found out AI was involved here, would they feel misled?

Higher stakes: content meant to represent your personal voice or expertise, such as thought leadership or opinion pieces. Client deliverables where someone is paying partly for human judgment. Content touching regulated areas like health, legal, or financial topics, where a reader’s expectation of professional human authorship is strong.

Lower stakes: internal tools. Grammar and spell-checking on content you wrote. Formatting assistance. A first draft you substantially rewrote with your own knowledge, specifics, and judgment. These uses are closer to editing than authoring, and most disclosure frameworks treat them differently.

The harder middle is where most businesses end up making inconsistent decisions. A post where AI wrote 60 percent of the structure and you filled in the specifics. A product description AI phrased but you reviewed. A FAQ where you prompted AI for questions and edited the answers. The most defensible approach is picking a consistent internal standard (if AI contributed more than a certain threshold, disclose it) and applying it the same way every time, rather than making case-by-case calls that seem arbitrary in hindsight.

This connects to the question that comes up in deciding what to automate with AI in the first place: the businesses that benefit most are not the ones who automate everything, but the ones who decide deliberately where the human stays in the loop. Disclosure is what makes that decision visible to customers.

What your disclosure policy signals

Here is the part most guides skip. The conversation about AI disclosure focuses almost entirely on the risk of disclosing (will it hurt trust, will it look unprofessional) and very little on what staying quiet signals.

The risk of not disclosing is not just legal. Customers who are already AI-aware may form their own judgment when they notice, rather than when they are told. That judgment is harder to manage because it is unsolicited. They figured it out, which means you were not forthcoming.

Businesses getting this right are treating disclosure as a position rather than a compliance checkbox. They are saying: we use AI tools, we use them well, and everything that goes out under our name has gone through our judgment first. That is a more credible position in 2026 than either “we do not use AI” (often not true and increasingly obvious) or silence (which asks customers to assume rather than know).

AI hallucinations are a real business risk, and a disclosure that includes human review is not just a trust signal. It is the actual differentiator between AI-generated and AI-assisted content. Telling customers you reviewed it makes the implicit claim that it is reliable. In a client relationship, that is the claim that matters most.

Understanding what happens to your data when you use tools like ChatGPT is the related question worth working through at the same time. Once you know where the data goes, your disclosure policy often starts writing itself.

Frequently asked questions

Does the law require you to disclose AI-generated content?

It depends on context. No universal US law requires a disclosure label on blog posts or emails. The FTC’s endorsement rules require disclosure when AI plays a material role in sponsored content. New York and the EU AI Act (both 2026) added specifics. The legal floor varies by context; the business case for transparency is often stronger than the minimum.

Do you need to disclose AI use in client work or deliverables?

If you are delivering professional analysis, a report, or a proposal, check your contract first. Many service agreements now include AI-use clauses. When in doubt, a brief note in the deliverable is safer than silence. A client who discovers that AI produced something they paid for as professional judgment will be more frustrated than one who was told upfront.

What should an AI content disclosure actually say?

Keep it brief and specific. “This article was drafted with AI assistance and reviewed by our team” works well for blog content. For advertising, FTC guidelines require disclosures to be clear and prominent. Vague language like “AI-enhanced” or “written with smart tools” does not tell readers enough. Specific beats vague.

Will disclosing AI hurt your credibility or search rankings?

The evidence points the other way. AI engines cite sources with strong E-E-A-T signals, and a transparent disclosure paired with visible human review tends to strengthen that signal. The credibility risk comes from customers who feel misled, not from customers who feel informed.

Is using AI for a first draft the same as AI-generated content?

Most guidelines draw a line between AI-assisted and AI-generated. Grammar fixes and phrasing suggestions generally do not trigger disclosure. A draft AI wrote and a human substantially rewrote sits in the middle. The clearest test: how much of the judgment, voice, and fact-checking did a human contribute? The more of that is genuinely human, the less this qualifies as AI-generated content.

---

The disclosure decision itself is the easy part. Most businesses can settle it in an afternoon: check what their contracts say, pick a standard, write a two-line template for the footer.

The harder question is what your disclosure policy reveals about the process behind it. Do you have a consistent internal standard for when AI is in the loop and when it is not? Does everyone on your team know it? Is the human review step actually happening, or is it something everyone assumes is happening?

Those are the questions your disclosure will keep surfacing, one published piece at a time. The gap between having a policy on paper and having one your team follows consistently (without it becoming extra overhead) is where most businesses end up stuck. That is also the part worth thinking through before you publish the next fifty pieces, not after.

If you are using AI in your content and operations and want that work to reflect something you would be proud to put your name on, that conversation is worth having before the next piece goes out. The free 30-minute call with VK, our AWS Certified Solutions Architect at our Castle Rock, Colorado studio, is where that conversation usually starts. Book it here or see how we approach Content & SEO and AI Automation services.