A Facial Recognition Vendor Told Hundreds of Stores It Had Zero Bias. It Had No Evidence.
What happened
IntelliVision Technologies Corp. sold its facial recognition software to retailers on a specific promise: the system ranked among the most accurate on the market, and it performed without racial or gender bias. Hundreds of stores deployed it on that basis. The Federal Trade Commission reviewed those claims in December 2024 and found that the company had no evidence to support either of them.
The FTC's complaint, voted through unanimously 5-0, alleged that IntelliVision made false, misleading, or unsubstantiated representations about the technology's accuracy rate and about its performance across individuals with different genders, ethnicities, and skin tones. A separate claim about the software's ability to detect spoofing, a technique used to deceive facial recognition systems with photographs or masks, was also found unsupported. The Commission's statement was direct: companies should not be touting bias-free AI systems unless they can actually back up those claims.
Under the proposed consent order settling the allegations, IntelliVision is prohibited from making misrepresentations about the accuracy or efficacy of its facial recognition technology, its comparative performance across demographic groups, and its spoofing-detection capability. The order also bars the company from making any representation about the effectiveness, accuracy, or lack of bias of its technology unless it can substantiate the claim before making it. The prohibition covers future marketing, not just the statements already at issue.
The scale of the deployment matters here. Facial recognition running across hundreds of retail locations affects shoppers who never consented to being scanned and who had no way to know the accuracy claims underpinning the system were unverified. A vendor's marketing language about "zero bias" does not stay inside a press release; it shapes how store operators configure alerts, how much weight loss-prevention staff place on a match, and what recourse a wrongly flagged person is assumed to need. If the underlying claim is unsupported, every downstream decision built on it inherits that same gap.
The FTC's action exposes a structural weakness that runs wider than any single vendor. Buyers of facial recognition software have no independent mechanism to verify performance claims before deployment, and the people scanned by these systems have even less access to that verification. A provable record of what a system actually did across demographic groups, retained and auditable alongside the marketing claims attached to it, would have surfaced the gap between IntelliVision's representations and its evidence long before the agency filed a complaint. Without that kind of record, unsubstantiated accuracy claims are indistinguishable from accurate ones until someone with enforcement power decides to check.
Reported impact
- Affected parties
- Not publicly disclosed
- Harm type
- Not publicly disclosed
- Scale
- Not publicly disclosed
- Financial impact
- Not publicly disclosed
- Regulatory action
- Not publicly disclosed
Classification
Relevant governance controls
Governance control mapping is not available for this record.
- No controls mapped
Not publicly disclosed
Control mapping is analytical. It does not state that any control would have prevented the incident.
Sources and evidence
This record was researched and written by the Index. The event is also catalogued in the following database, which is listed for cross-reference.