On December 3, 2024, the FTC took action against IntelliVision Technologies Corp., a developer of AI-powered facial recognition software for home security and smart devices, for alleged deceptive marketing practices. This action continues the agency's aggressive posture towards marketing of AI-related technologies that reached a fever pitch with Operation AI Comply in September. The case also represents the latest chapter in the agency's scrutiny of facial recognition technology ("FRT"), increased in recent years through settlements with EverAlbum (2021) and Rite Aid (2023).
Claims v. Reality
IntelliVision's marketing materials made several bold assertions about its software:
One of the highest accuracy rates in the industry.
Training on millions of diverse faces.
Equal detection capability across all demographics.
Robust anti-spoofing technology to identify genuine users.
However, the FTC's investigation painted a different picture. The company had only used approximately 100,000 faces but used variations to make the data set appear larger. Its software exhibited differing error rates across demographics and fell outside the top 100 most accurate facial recognition software according to National Institute of Standards and Technology (“NIST”) testing. The anti-spoofing technology lacked sufficient testing to validate accuracy claims, particularly for skin tone and other demographic characteristics.
The FTC’s order included a prohibition on misrepresentations about its technologies, a substantiation requirement for claims through “competent and reliable testing,” and documentation and compliance requirements.
The FTC demonstrates its concern about the use of certain technologies, like connected devices or SDKs, through repeated enforcement actions and their companion blog posts. AI has naturally been within the agency’s focus, whether the hook is unfairness as in Rite Aid or deception in Operation AI Comply. Add FRT to the list with this case, with warning to businesses that this is unlikely to be the final action or investigation involving FRT or inflated AI marketing claims.
Practical Guidance for Navigating Regulatory Expectations
Be precise with ambiguous terms. While the vote was 5-0 — indicating momentum that will continue into the new administration — Commissioner Ferguson released a concurring statement taking umbrage with IntelliVision’s claim that its technology was free of “bias.” The company failed to define “bias,” a notoriously multifaceted term, leading Ferguson to declare the company had “the burden of substantiating all reasonable interpretations that consumers may have given its claim that its software had ‘zero gender or racial bias.’” His remark references a central tenet of Section 5 cases from the 1984 policy statement on deception that deserves resurfacing for marketers in the age of AI: “When a seller’s representation conveys more than one meaning to reasonable consumers, one of which is false, the seller is liable for the misleading interpretation.”
Test, test, test — and keep the receipts. Each claim made — in marketing materials or otherwise — should be directly connected to documented testing by qualified professionals. It should be industry-standard and tailored to the claim; anti-spoofing claims should be supported by testing verifying the technology can differentiate between a live person and a photo or video.
Evaluate your technology with industry-standard testing, like NIST’s Facial Recognition Vendor Test. Align your claims with those results, particularly if they are publicly disclosed. NIST has similar programs for speech, fingerprint, and iris recognition, iBeta for liveness detection, and the FIDO Alliance for remote biometric identity verification. There’s a testing process for your technology, too.
Looking Ahead
This case makes clear the FTC expects precise, well-documented marketing claims about AI capabilities, particularly when marketing sensitive technologies like facial recognition. Companies should expect continued scrutiny of AI marketing claims, especially those touching on accuracy, bias, or demographic performance. As the regulatory landscape evolves, the safest path forward is through rigorous testing, careful documentation, and measured marketing claims.
