Published Date : 03/12/2024Â
The U.S.
Federal Trade Commission (FTC) has taken action against IntelliVision, a company that specializes in face biometrics and liveness detection software.
The FTC has issued a consent order prohibiting IntelliVision from making misleading claims about the accuracy, lack of bias, and spoof protection capabilities of its products.
IntelliVision had claimed that its facial recognition technology was free from gender and racial bias and was one of the most accurate on the market.
However, the FTC found that these claims were not supported by evidence.
The company also claimed to have trained its biometric algorithms on millions of faces, but in reality, it had used images of approximately 100,000 people and created variants of those images using technology.
Additionally, IntelliVision's presentation attack detection technology, which was marketed to protect against spoofing with photos or videos, was found to lack support.
The FTC has now banned IntelliVision from making any claims about the accuracy, lack of bias, or spoof protection unless they are backed by 'competent and reliable testing.'
The issue is not that testing has not been conducted, but rather that the results of this testing, particularly by the National Institute of Standards and Technology (NIST), show significant discrepancies.
NIST's testing revealed that matching rates vary between different demographics, and as of December 2023, IntelliVision's software did not show a false non-match rate within the top 100 in the FRTE 1 1 category.
In 2022, IntelliVision had submitted an algorithm to NIST’s FRTE 1 N, which currently ranks 74th in Border ΔT ≥ 10 YRS and 95th in the Visa-Kiosk category.
The FTC began observing discrepancies between IntelliVision's claims and the facts in late 2018, shortly after the company was acquired by Nortek Security & Control.
IntelliVision's facial recognition technology is used in various smart home systems, including panels sold by Nice North America, as noted by the FTC.
The Commission voted unanimously to issue the complaint and accept IntelliVision's settlement agreement.
Each violation of the agreement will result in a civil penalty of up to $51,744.
The FTC has also published a biometrics policy statement and a blog post with tips for biometrics providers on how to 'avoid breaking the law and misleading consumers.'
In 2022, the FTC's biometrics policy statement focused mainly on data protection issues.
However, a later action against Rite Aid addressed both data protection and the effectiveness of the software.
The FTC's actions underscore the importance of transparency and accuracy in the marketing of biometric technologies, ensuring that consumers are not misled by false or unsupported claims.Â
Q: What is the main issue with IntelliVision's claims?
A: The main issue with IntelliVision's claims is that they were not supported by reliable evidence. The company claimed its facial recognition technology was free from bias and highly accurate, but testing by the National Institute of Standards and Technology (NIST) showed significant discrepancies.
Q: What action did the FTC take against IntelliVision?
A: The FTC issued a consent order prohibiting IntelliVision from making any claims about the accuracy, lack of bias, or spoof protection of its biometric technology unless backed by 'competent and reliable testing.'
Q: When did the FTC start observing discrepancies in IntelliVision's claims?
A: The FTC began observing discrepancies between IntelliVision's claims and the facts in late 2018, shortly after the company was acquired by Nortek Security & Control.
Q: What are the penalties for violating the FTC's consent order?
A: Each violation of the FTC's consent order will result in a civil penalty of up to $51,744.
Q: Who uses IntelliVision's facial recognition technology?
A: IntelliVision's facial recognition technology is used in various smart home systems, including panels sold by Nice North America.Â