Why “AI-washing” is now a brand risk

“AI-powered” has become one of the most common claims in marketing. When the product behind the label does little more than before, customers notice, and so do journalists and regulators.

The term for this is “AI-washing”: overstating how much AI a product uses, or what it can do. It is the AI version of greenwashing, and since 2024 it has moved from a reputation problem to a legal one.

The risk

Overstated AI claims erode trust quickly. A single feature that doesn't live up to its label can color how people read everything else you say.

The trust problem shows up even when claims are honest. A Washington State University study of more than 1,000 U.S. adults found that products described as using artificial intelligence were less popular. Lead author Mesut Cicek said: “When AI is mentioned, it tends to lower emotional trust, which in turn decreases purchase intentions.”¹ The effect was stronger for high-risk purchases, such as expensive electronics, medical devices, or financial services.¹

So the label carries a cost even when it is true. When it is not true, the cost is much higher.

What regulators have done

In the U.S., two regulators have led the way.

The Securities and Exchange Commission (SEC) brought its first AI-washing cases in March 2024 against two investment advisers, Delphia and Global Predictions, which agreed to pay $400,000 in total civil penalties. According to the SEC, Delphia claimed it used client data to make its artificial intelligence smarter, which the agency found it could not do, and Global Predictions falsely claimed to be the “first regulated AI financial advisor.” The SEC’s then chair, Gary Gensler, said: “Investment advisers should not mislead the public by saying they are using an AI model when they are not. Such AI washing hurts investors.”²

The SEC’s enforcement director, Gurbir S. Grewal, spelled out the lesson: “if you claim to use AI in your investment processes, you need to ensure that your representations are not false or misleading.”² The same release warned that public companies making claims about their AI adoption must “remain vigilant about similar misstatements.”²

More cases followed. In January 2025, the SEC charged Presto Automation, a restaurant technology company, over misleading claims about its drive-thru AI product.³ In April 2025, it charged Albert Saniger, founder of the shopping app Nate, alleging that he raised over $42 million from investors while claiming the app used AI to complete purchases, when contractors were handling orders by hand.⁴

The Federal Trade Commission (FTC) launched “Operation AI Comply” in September 2024, with actions against companies including DoNotPay, which claimed to offer “the world’s first robot lawyer.” The FTC’s then chair, Lina M. Khan, said: “Using AI tools to trick, mislead, or defraud people is illegal. The FTC’s enforcement actions make clear that there is no AI exemption from the laws on the books.”⁵

In August 2025, the FTC finalized an order against Workado, which had marketed its AI content detector as suitable for everyday users. The FTC’s complaint alleged that the model behind it was trained or fine-tuned to effectively classify only academic content. Workado may no longer make accuracy claims unless it has competent and reliable evidence to support them.⁶

In Europe, the AI Act adds transparency duties. Its rules for telling people when they are dealing with a chatbot, and for labeling deepfakes, came into effect in August 2026.⁷ That is not an AI-washing rule as such, but it raises the general expectation that companies say clearly what their AI does.

These are mostly U.S. cases about investors and consumers, but the principle travels. Advertising rules in most countries already require claims to be truthful and backed by evidence, and an AI claim is simply another claim.

Coming soon: The EU AI Act for marketing teams →

The pattern behind the cases

The cases share a few patterns worth checking against your own marketing:

• Claiming AI that isn’t there. Saying a product uses AI when people or simple rules do the work.
• Overstating what it does. Saying the AI “eliminates” a task when people still step in for most of it.
• Unproven numbers. Accuracy or performance claims without testing to back them.
• Borrowed capability. Presenting a third party’s technology as your own.

None of these rules is new. Truth in advertising applies to AI claims the same way it applies to “organic” or “clinically proven.” What is new is how often AI is claimed, and how closely regulators are now reading those claims.

A better approach

Describe the benefit and the task, not the technology. “Drafts your weekly report in two minutes” is more credible, and more useful, than “AI-powered reporting.”

That approach solves both problems at once. It avoids the trust penalty the WSU study found, because it leads with what the customer gets. And it is easy to substantiate, because a specific benefit can be tested and shown.

Where AI is a real differentiator, explain it in plain words: what it does, what a person still checks, and what its limits are. Being honest about limits builds more trust than a big claim does.

The WSU researchers tested this across many kinds of products. As Cicek put it: “We tested the effect across eight different product and service categories, and the results were all the same: it’s a disadvantage to include those kinds of terms in the product descriptions.”¹ His advice to marketers was to focus on describing the features or benefits and to avoid the AI buzzwords.¹

Five questions before you say “AI”

1. What exactly does the AI do, and what do people still do?
2. Could we show a customer or regulator the evidence for this claim today?
3. Is the number we quote based on our own testing, in conditions like the customer’s?
4. Is the technology ours, or a supplier’s, and do we say so honestly?
5. Would the claim still sell the product if we removed the word “AI”?

If the answer to the last question is no, the product may need work, not the copy.

Run the questions past legal or compliance for anything that will appear in investor materials, pricing pages, or regulated sectors, where the bar for evidence is highest.