Cox Media Sold Clients a Surveillance Tool That Never Existed. The FTC Finally Noticed.

Cox Media Group called it "Active Listening." The pitch deck, still visible on DocumentCloud, described a product that used artificial intelligence to harvest what people were actually saying — in real time, through their phones and smart speakers — and sell it to advertisers. "Smart devices capture real-time intent data by listening to our conversations," the deck stated. "Advertisers can pair this voice-data with behavioral data to target in-market consumers."

It was, according to the Federal Trade Commission, a lie.

On May 21, the FTC announced a $930,000 settlement with CMG and two smaller companies, MindSift and 1010 Digital Works, resolving charges that they deceived customers about what the service actually did. CMG, the lead defendant, will pay $880,000. The two vendors each pay $25,000. The consent order is subject to a 30-day public comment period before becoming final.

The case turns on a document that is still available online: CMG's own sales pitch, uploaded to DocumentCloud in 2023. It makes specific technical claims that the FTC has now officially weighed and rejected.

What the deck promised

The pitch deck, first reported by 404 Media in December 2023, described a service built on a simple premise: smart device microphones are always listening, and that audio can be captured, processed by AI, and sold to advertisers hungry for signals about what consumers actually want.

CMG made at least four specific claims the FTC later called out:

"Smart devices capture real-time intent data by listening to our conversations." The pitch deck did not describe this as a metaphor or an analogy. It stated it as a technical fact about how the service worked.

"Advertisers can pair this voice-data with behavioral data to target in-market consumers." The deck claimed the service not only captured voice data but could combine it with behavioral profiling to create a targeting system. The FTC found no voice data was involved at any stage.

"We use AI to collect this data from 470+ sources." The deck cited a specific number — 470+ — and attributed it to an AI-powered data collection pipeline. The FTC found the service used no AI and sourced nothing from those claimed channels.

"You reach your potential customers before your competitors." The pitch deck promised competitive advantage: that CMG's data would put clients ahead of rivals in the market for the same consumers. The FTC found the ad placement was not accurate in customers' desired locations.

That is a list of four specific, verifiable promises. The FTC found each one false.

What the FTC found

According to the agency's complaint, the "Active Listening" service did not listen to anyone. It did not process voice data. It did not use artificial intelligence in any substantively meaningful way. What it actually did was buy email lists from other data brokers and resell them at a markup — the same commodity data brokerage that has underpinned decades of digital advertising, dressed in science-fiction clothing.

The consent claims were also fictional. CMG argued that consumers had consented to the surveillance by clicking through app terms of service — a mechanism the FTC squarely rejected. "Clicking through mandatory terms of service does not constitute opt-in consent," the agency stated.

The companies were charged under FTC Act Section 5, which prohibits unfair or deceptive acts or practices.

The penalty math

The $930,000 settlement sounds significant until you read further into the consent order. Once final, each violation carries a civil penalty of up to $53,088. That figure is not theoretical — it is the actual exposure if CMG violates the order's terms, which prohibit the company from making the same claims again.

The FTC Commission voted 2-0 to issue the proposed administrative complaints and accept the consent agreements.

The deflection

CMG has a response ready. "Our local marketing team relied on marketing materials provided to us by a third-party vendor about their product," the company told WIRED. That framing — a rogue vendor misled an otherwise innocent local franchise — is the first line of defense in almost every case like this. Whether it holds up depends on what the FTC's investigation found about corporate knowledge and approval chains, which the settlement documents do not fully resolve.

Senator Marsha Blackburn opened a Senate probe of Big Tech platforms following CMG's initial admission in September 2024, adding congressional pressure to the regulatory enforcement track. CMG is also a Facebook Meta Audience Network partner — a relationship that gave the company credibility in the ad-tech ecosystem and may have helped it sell the product to buyers who should have asked harder questions.

Why this matters beyond the settlement

The FTC has now drawn a line, however small the fine, on a specific form of AI-washing: the claim that a product listens to consumer conversations in real time using smart device microphones. That claim is not merely puffery — it is a specific technical assertion about how data is collected and processed. When a company makes that claim and the product does something completely different, it is deception, not aggressive marketing.

The broader lesson is narrower than the conspiracy theorists think. The conspiracy theory — that tech companies are secretly listening to your phone — turns out to be false. But the actual threat, quieter and less cinematic, is that companies are selling the ordinary data broker surveillance economy at a premium by wrapping it in the language of that conspiracy theory. The FTC caught one company doing it. The harder question is how many others are doing it quietly, without a congressional hearing or a viral pitch deck.

The consent order's 30-day public comment period is open. Comments will be reviewed before the order becomes final.