The Federal Trade Commission just did something every small business owner should notice, even if you never plan to sell an AI product.
On July 1, 2026, the FTC opened public comment on a proposed policy statement about AI accuracy. The agency is asking how companies describe what their systems do and whether they are shaping outputs in ways that make them look more trustworthy, precise, or capable than they really are. FTC legal library
That is not the same thing as a new rule.
It is a signal.
And the signal is simple: if you are making claims about AI, the FTC wants to know whether you can back them up.
Why This Matters Outside Big Tech
Most small business owners read FTC news and assume it is for the giants.
Not this one.
If you market anything with AI in the description, the question is no longer just whether the software works in your demo. The question is whether your claims survive real use:
- Does the tool actually improve accuracy, or just sound impressive in a pitch deck?
- Are you saying "automated" when a human is still cleaning up the output?
- Are you promising "instant" results that still require three review steps?
- Are your testimonials and before-and-after claims representative, or the best possible case?
If those questions make you wince, good. That is the point.
The FTC's AI page already shows where this can go. The agency has been active on cases involving misleading AI claims, deceptive marketing, and unsupported promises about what AI can do. FTC Artificial Intelligence page
The Owner Takeaway
This is the practical version of the story:
If your business uses AI in customer-facing work, you need a paper trail that answers three questions.
- What exactly does the tool do?
- What still needs human review?
- What proof do you have for the claims you make about it?
That applies to marketing copy, sales demos, onboarding language, support chat, hiring tools, and anything else a customer or regulator might later read.
If a vendor says its AI is "more accurate," ask what that means.
More accurate than what?
Measured how?
On what dataset?
Under what conditions?
If you cannot answer those questions in plain English, you probably do not have a claim. You have a slogan.
What To Do This Week
If you run a small business, here is the quick audit I would do before the month ends:
- Pull every page where you mention AI and flag every promise word: accurate, instant, guaranteed, autonomous, fully automated, human-level, or similar.
- Check whether those promises are supported by something real, not just a vendor webinar.
- Make sure customer-facing workflows still have a human review step where mistakes would cost money, trust, or legal trouble.
- If a vendor is giving you a sales pitch, ask for examples from actual use, not just screenshots.
The boring part is the point.
The FTC is not asking small businesses to stop using AI.
It is asking them to stop pretending a polished demo is the same thing as a defensible claim.
That is good news for owners who care about honesty, because it rewards the businesses that already tell the truth about what their tools can and cannot do.
The rest of the market is about to get more expensive to bluff.
Sources: FTC legal library - Federal Trade Commission’s Proposed Policy Statement Concerning the Suppression of Accuracy In Artificial Intelligence Systems; FTC Artificial Intelligence page