IntelliVision Facial Recognition: Bias-Free and Accuracy Claim Evidence Questions
The FTC finalized an order against IntelliVision over claims that its AI-powered facial recognition software had high accuracy, zero gender or racial bias, and anti-spoofing capability. This case shows the evidence burden behind demographic performance and safety claims.
Source update, company response, or correction? Send a private note for review →
What was claimed
IntelliVision's facial recognition software was marketed with claims about high market accuracy, zero gender or racial bias, training on millions of faces, and anti-spoofing technology that could prevent the system from being fooled by a photo or video image.
- Risk pattern
- Bias-free and anti-spoofing AI claim without subgroup testing, training-data support, or attack-scenario limits
Why this mattered
Bias-free and anti-spoofing claims create a high evidence burden because buyers need subgroup performance, error-rate spread, training-data documentation, deployment limits, and test results for known spoofing methods. A general accuracy claim does not show whether performance holds across demographic groups or adversarial conditions.
What the source said
The FTC said IntelliVision lacked evidence for claims that the software had one of the highest accuracy rates on the market, performed with zero gender or racial bias, was trained on millions of faces, and had adequate anti-spoofing support. The final order restricts misrepresentations about accuracy, efficacy, demographic performance, and anti-spoofing capability.
Evidence gap / buyer questions
Subgroup performance by gender, ethnicity, and skin-tone groups; sample size and training-data documentation; benchmark design; error rates by deployment setting; anti-spoofing test methods; attack types tested; and limits for image quality, lighting, angle, and camera conditions.
- What are the false positive and false negative rates across gender, ethnicity, and skin-tone groups?
- What training data supports the claim, and how was the dataset audited for coverage gaps?
- Which spoofing methods were tested, and which attack types remain outside the claim?
- Does the claimed accuracy apply to our camera quality, lighting, image angle, and user population?
How this applies to your vendor evaluation
If a vendor you are evaluating makes a claim with this pattern, copy the exact sentence and review that wording against the evidence standard this case documents.
Wording boundary direction
Tested on [named dataset] across [listed demographic groups] under [conditions]; subgroup error rates and anti-spoofing limitations are available at [link].
A lower-risk wording boundary narrows the scope, discloses the test conditions, and does not overstate what is covered.
Update and response status
Disclaimer / correction note
This case description draws from the FTC source cited above. It is not legal advice, a biometric technology audit, or a determination that any product is safe or suitable for deployment.
This tool generates evidence-burden notes, evidence requests, and buyer questions based on publicly accessible source content. It does not determine whether a product is true, false, compliant, or suitable for any purpose. It is not legal, investment, procurement, or professional compliance advice. See the full disclaimer.
Source update, company response, or correction? Submit a correction or source note →
Check a similar public AI claim
Review the exact public URL or wording for evidence gaps and buyer questions.