FTC v. IntelliVision: Companies Need Evidence for ‘Bias-Free AI’ Claims

The enforcement action targeted unsupported accuracy and zero-bias marketing claims.

What happened—and what the record establishes

The FTC alleged that IntelliVision Technologies lacked evidence for claims that its facial-recognition software had very high accuracy and zero gender or racial bias. The Commission finalized an order in 2025 restricting unsubstantiated representations.

The case is about marketing substantiation, not a finding that every use of the software injured a specific person. It shows that AI vendors must possess competent and reliable testing before making objective performance claims.

Customers purchasing biometric tools should ask for test populations, false-positive and false-negative rates, subgroup results, operating conditions, spoofing methodology, and independent validation—not accept a single headline percentage.

Questions the law must answer

• What test supports the advertised claim?
• Does the test population match the deployment population?
• Are subgroup and real-world error rates disclosed?

Practical takeaways

• ‘Bias-free’ is a factual claim that needs proof.
• Laboratory accuracy may not predict field performance.
• Procurement contracts should require documentation and incident cooperation.