AI Claims — Yay or Oy? A Recent Warning Letter Tells a Cautionary Tale

A recent Food and Drug Administration Warning Letter offers a timely reminder that how a digital health product is promoted can be just as consequential as how it functions. An August 2025 Warning Letter, issued after an FDA inspection, identified promotional claims for a software-based health product that, in the agency’s view, crossed into diagnostic territory — largely due to references to artificial intelligence.

While the Warning Letter addressed a single company, the broader message applies across the growing field of AI-enabled health tools: promotional claims define intended use, and words such as detect, predict, or screen can quickly shift a product into device oversight — whether or not that was the company’s intent.

Promotional Claims Determine Regulatory Boundaries

FDA focused on marketing statements including exercise and rehabilitation use suggesting that the software could assess or predict health risks such as fall propensity or cognitive decline.¹ Even without explicit references to diagnosis, the agency found that these claims implied clinical evaluation and decision-support functionality beyond the product’s 510(k)-cleared purpose.

Key points from FDA’s position include:

• Intended use is shaped by promotion, not only labeling.
• Language matters — words implying clinical or predictive functionality can transform a wellness product into a medical device.
• Overbroad claims may cause an otherwise low-risk or previously cleared product to fall back under stricter device oversight, even if its technical features have not changed.
• Predictive AI claims draw attention when they imply the ability to diagnose, detect, or mitigate disease.

In short, a company can have a valid 510(k) clearance, but if its marketing suggests new or expanded clinical functionality, those statements can effectively move the product outside the scope of its clearance — and back into regulatory hot water.

Caution on References to FDA Status

FDA also reiterated its sensitivity to promotional references implying formal regulatory approval or endorsement. To stay compliant, companies should:

• Avoid phrases such as “FDA approved,” “FDA certified,” or “FDA registered” that can imply FDA preference or higher-level review (e.g., approved under a Premarket Approval Application).
• Ensure marketing materials, investor decks, and app store listings accurately reflect the product’s current regulatory status.

Misuse of FDA references, even unintentionally, can mislead consumers and invite enforcement.

Building a Strong Internal Review Process

Although FDA’s primary focus in this case was on promotional claims, it also identified deficiencies in quality system and complaint-handling procedures. This reinforces that marketing compliance and operational readiness go hand in hand.

• A sound compliance infrastructure should include cross-functional claim review involving regulatory, legal, and marketing teams.
• Documented processes to assess whether new features, algorithms, or analytics (particularly those involving AI) affect regulatory status.
• Regular audits of promotional materials to ensure consistency with the current indication.

The Takeaway – AI: The Good, the Bad, and the Scrutinized

AI continues to bring remarkable opportunities to healthcare innovation —but also heightened regulatory attention. As this and other recent FDA enforcement actions show (including its recent AI-related Warning Letter), the same technology that drives advancement can also amplify compliance risk when claims outpace clearance. In other words: with the good comes the guarded.

[1] FDA also noted concerns with the company’s compliance with Quality System Regulation and Medical Device Reporting Violations in the Warning Letter. We will only focus on the promotional claims in this Bulletin.