NAD’s AI-washing cases remind advertisers to qualify performance claims and support comparative claims

In a March 2026 client alert on advertising dispute trends, we discussed BBB National Programs’ National Advertising Division’s (NAD) focus on substantiation for performance and superiority claims, particularly for AI-enabled and health-related products. Recent NAD decisions cover those focus areas, complement several recent Federal Trade Commission (FTC) “AI-washing” cases, and provide a good reminder generally about the legal standards for unqualified and comparative advertising claims.

Starkey Laboratories

On September 8, 2026, NAD issued a final decision in a case involving Starkey Laboratories, Inc., arising out of a competitor challenge. NAD’s inquiry centered on claims for the company’s Omega AI hearing aids, described as using “Deep Neural Network” technology. Specifically, NAD reviewed the company’s substantiation for claims that the product provided “better hearing all around,” better spatial awareness, and superior performance to competitors, and that Starkey is “the leader in hearing healthcare.” NAD recommended that Starkey modify or discontinue those claims, and the company agreed to do so.

NAD noted first that “[w]hen emerging science is the basis for a claim, it must be carefully qualified and include any significant limitations or inconsistencies in the evidence.” In another recent decision, using a line similar to other FTC warnings, NAD stated that “[c]ompanies eager to tout their AI technology must ensure that such claims are truthful and non-misleading, especially when consumers cannot fully evaluate AI claims for themselves.” Citing FTC precedent, NAD has also stated that claims for health-related products must be “supported by science which often includes methodologically sound testing.”

In Starkey, NAD found the challenged claims generally overbroad and not properly qualified relative to the supporting evidence. The broad “better hearing” claim was based on a consumer study in which over 40 percent of respondents gave no hearing aid preference, above NAD’s 20-percent no-preference threshold for unqualified preference claims. As for the broad spatial awareness claim, NAD highlighted that comparative claims should make clear whether the comparisons are to competitors or to one’s own prior technology. NAD also stated that the company could not support claims that it is “the leader in hearing healthcare,” rather than “a” leader, and that leadership claims should be tied to clear metrics.

The decision also features an interesting discussion of Starkey’s use of an AI tool both to test an AI product and to use the resulting data to substantiate marketing claims. In discussing the company’s claims of superior performance, NAD noted that AI testing is “an emerging area with potential benefits” but said that companies need to validate such testing for the specific claims at issue, “particularly when claims convey real-world consumer benefits.”

Other recent NAD decisions

NAD’s other recent AI-washing decisions have had varied outcomes, with its close reviews resulting in recommendations that a company modify or discontinue certain claims, while finding adequate support for other challenged claims.

For example, it has distinguished a broad claim that an AI-enabled product is “proven” to have the desired outcome from a narrower claim that it had this outcome in a particular community and to a specifically calculated percentage. In one case, it found the former claim unsupported and the latter claim adequately qualified and substantiated. In another case, it said that a company should modify claims about an AI-powered product’s ability to detect emotion and clearly and conspicuously disclose the product’s “functionality limitations.”

NAD has also considered “industry-leading” or other superiority claims in several matters involving AI-enabled or “smart” products. As in Starkey, NAD has explained that unqualified comparative claims, including leadership or category-wide superiority claims, must be supported by reliable evidence demonstrating performance at or above that of competing products on “meaningful industry metrics.”

These findings are notable because some may view simple “industry-leading” claims as puffery, similar to a statement that one’s product is “the best.” NAD may sometimes use other logic, however, finding in one case that a company calling its product the “smartest” in its category had adequate support because it had “smart” features that were “unique” in the industry.

With Starkey as the latest example, most of NAD’s AI-washing decisions over the last two years have focused on claims in the healthcare industry, with multiple matters addressing AI-powered products designed to help care for seniors and children and keep them safe. These are evergreen concerns for NAD, and these claims are expected to remain a focus.

FTC comparisons

As explored in past client alerts here and here, the FTC has been a leading enforcer against allegedly deceptive AI claims. Like the NAD cases, a few of the FTC's recent, AI-focused cases featured unqualified and comparative claims. In both Workado and accessiBe, the FTC addressed allegedly overbroad performance claims. In Workado, involving an AI writing detection tool, the FTC alleged that the company’s accuracy claims should have been qualified to reflect that they were substantiated only for academic writing. In its action against accessiBe, involving a website accessibility tool, the FTC alleged that the company failed to disclose limits on the tool’s ability to make full websites accessible.

The FTC has also focused on comparative claims in its AI cases, including Evolv Technologies and IntelliVision Technologies. In Evolv Technologies, the agency alleged that the company made unsupported claims that favorably compared its security screening product to traditional metal detectors. In IntelliVision Technologies, the FTC alleged that the company lacked substantiation for its comparative claim that its facial recognition software had one of the highest accuracy rates on the market.

Takeaways

Beyond their AI-specific findings, Starkey and other NAD AI-washing decisions serve as reminders of broader advertising principles that, if overlooked, could expose companies to competitor challenges and regulatory scrutiny.

• A company’s substantiation for an advertising claim, especially one involving scientific or health-related statements, must properly match the specific claim and reflect real-world conditions. That point may be especially relevant as more companies use AI-based testing in place of testing with human participants. Recent decisions indicate that NAD closely analyzes supporting evidence.
  
  
• A heightened standard applies when advertisers use language like “proven,” and even high-quality evidence may not meet it. NAD may, for example, look at whether such evidence is sufficient to establish causal attribution. Under the “establishment claim” doctrine, when an advertiser affirmatively represents that its claims are backed by scientific proof, it must possess the specific level of proof that the claim conveys to consumers. For health and safety products, this generally requires well-controlled studies conducted on the product itself.
  
  
• NAD’s treatment of “industry-leading” claims reinforces that category-wide superiority claims require comparative proof, typically head-to-head testing against competitors on relevant industry benchmarks. NAD generally does not find such claims to be substantiated by evidence of the company’s own performance alone. This is a particularly active area for competitor challenges both before NAD and in federal court under the Lanham Act. On the other hand, some companies may be able to substantiate certain superiority claims, such as “smartest,” by demonstrating that a product has “unique” features.

Advertisers are encouraged to consider the following measures:

• Confirm adequate, real-world support for specific claims
  
  
• Ensure “proven” or “clinically tested” claims are supported by sufficient scientific evidence, not only observational data
  
  
• Ensure superiority claims are narrowly drawn to match available comparative evidence

Whether for AI-related or any other claims, regulators and competitors may challenge gaps between advertising language and supporting evidence.

For more information, please contact the authors.