Workado AI detector: the FTC order behind accuracy claims

What the public record shows

The FTC’s Content at Scale AI case timeline identifies Workado as the respondent and records approval of a final consent order on August 28, 2025. The order requires competent and reliable evidence for advertised AI-detection accuracy. The case summary describes the regulator’s allegations about unsupported accuracy claims.

Where the promise broke

An accuracy percentage is incomplete without the dataset, decision threshold and treatment of mistakes. A detector can look persuasive on selected examples while producing costly false accusations on the text a buyer actually processes. Our evaluation approach keeps human-written and generated test samples separate, uses representative languages and documents false positives as well as missed detections. Do not accept a demonstration consisting only of examples supplied by the seller.

What this record does not establish

A consent order is distinct from a criminal conviction. The record establishes advertising restrictions and a substantiation requirement; it does not prove that every detector result is wrong. Its timeline describes a final order even though the general case-status field says pending. This article follows the dated action, with that distinction visible.

The buyer lesson

Request the evaluation protocol and test content you control. Decide what consequence a flag triggers and require human review before consequential decisions. Keep a way to challenge an incorrect result. Record sample origin, score, threshold, true outcome and reviewer decision; a blended accuracy figure can hide the failure mode that matters most to your team.

Sources: FTC case timeline and final order · 28 August 2025