Skip to content
Back to the Case Library
United States (Federal)Computer visionAI hardware & compute2024

Evolv pursued by the FTC over exaggerated AI screening claims; school customers allowed out of multi-year contracts

Subject: Evolv Technologies (Massachusetts-based AI security screening vendor)

Verified against official sources · 2026-09-12Verified against official text

What happened

Evolv Express is an AI-powered security scanner used in thousands of schools as well as venues such as stadiums and hospitals; schools make up half of its business, with Express scanners in over 800 schools across 40 states. Its marketing emphasised AI and positioned the product as a high-tech alternative to traditional metal detectors.

How it was handled

The FTC alleged Evolv misrepresented that Evolv Express would detect all weapons; would ignore harmless personal items without requiring people to remove them from pockets or bags; would detect weapons more accurately and faster than metal detectors; would reduce false alarm rates; and would cut labour costs by 70% compared with metal detectors. The complaint states the scanners failed in several instances to detect weapons in schools while flagging common items such as laptops, binders and water bottles: in October 2022 a seven-inch knife was brought into a school undetected and used to stab a student, after which officials raised the sensitivity setting, producing a 50% false alarm rate. In 2023 Evolv introduced a more sensitive setting aimed at detecting more knives, while stating that some knives would still be missed, more false alarms would occur and additional staffing might be required, and advising schools to add conveyor belts and other measures to divert harmless items by hand — which, the complaint says, makes the system more like a lower-cost traditional metal detector.

Outcome

The FTC filed its complaint and a stipulated order in the U.S. District Court for the District of Massachusetts (the order takes effect once signed by the judge). The proposed settlement requires Evolv to notify certain K-12 school customers that they may cancel contracts signed between 1 April 2022 and 30 June 2023, and prohibits misrepresentations about its products’ ability to detect weapons, to ignore harmless personal items, and to do so without requiring visitors to remove items from pockets or bags; about accuracy in detecting weapons and false alarm rates, including comparisons with traditional metal detectors; about screening speed compared with metal detectors; about labour costs, including such comparisons; about testing or the results of any testing; and about any material aspect of product performance, including the use of algorithms, artificial intelligence or other automated systems or tools.

What it means for Chinese exporters

  • · Performance claims for hardware plus AI are tested against real deployment behaviour, not lab metrics. When selling vision inspection, security or quality-control equipment abroad, state the test conditions (threshold settings, item classes, environment) and be explicit about the false-positive/false-negative trade-off.
  • · Business-benefit claims such as "cuts labour costs by X%" also need substantiation, and are easier to falsify than technical metrics — if customers in fact need more staff or added conveyor belts, the claim is deceptive.
  • · Products with adjustable sensitivity must disclose the trade-off. Writing "raising sensitivity reduces misses but increases false alarms" into product documentation and sales material, and confirming the customer’s chosen operating configuration in writing at delivery, is the single most effective protective step.
  • · The longer the contract in sensitive settings — schools, hospitals, public transport — the greater the risk. The FTC gave school customers a direct right to cancel; design pricing and contract structure with a controlled exit rather than relying on multi-year lock-in.
  • · One safety incident turns all your marketing material into evidence. Before launch, run a worst-case review of every external claim: assuming the product fails in the field and someone is hurt, which sentences in the current copy would be quoted back at you?

Sources (1)

Content as of 2026-09-12 · corpus 2026.09|Compiled from public official sources with links and review dates for compliance decision support. Rules change; have licensed professionals review before external disclosure, regulatory filing or contractual commitments.

GDPR · PIPL · CCPA control mapping
Traceable AI · cited when sourced
ISO 27001 alignment in progress
WCAG 2.1 AA accessibility target
Evolv pursued by the FTC over exaggerated AI screening claims; school customers allowed out of multi-year contracts · Public case | HaiWen HQ