I watched a slide deck promise the impossible: that your phone’s small talk could be sold to local businesses. You felt your privacy tilt—sudden and small, like a door left ajar. I remember thinking: if this is true, what else have we been selling without knowing?
The sales pitch looked ordinary — then it got creepy
Sales materials literally bragged that devices were “listening” and that advertisers could target people by phrases they spoke at home.
You’ve seen the banner: “It’s True. Your Devices Are Listening to You.” Cox Media Group ran that language in 2023, leaning into the paranoid thrill that keeps readers scrolling. The offering, dubbed Active Listening, promised ad targeting built from voice data captured on phones and smart devices — down to lines like, “The AC is on its last leg!”
But according to the Federal Trade Commission, the claim was fiction. The product did not harvest conversations. Instead, the companies repackaged and resold existing email lists from data brokers at a markup. The sales pitch was a wolf in sheep’s clothing.
A real enforcement move — and what the FTC actually said
Records and a press release from the FTC show investigators dug into marketing collateral, contracts, and vendor claims.
The FTC accused Cox Media Group, MindSift, and 1010 Digital Works of deceiving customers about the technology and about consumer consent. Christopher Mufarrige, Director of the FTC’s Bureau of Consumer Protection, framed it plainly: firms must be honest with customers. The agency added that clicking through app terms didn’t equal opt-in consent for an invasive service that would mine voice data from inside homes.
Under the proposed settlement, Cox will pay $880,000 (€810,000), while MindSift and 1010 Digital Works will each pay $25,000 (€23,000). The nearly $1 million total (€920,000) will go toward refunds for affected CMG customers.
Can companies legally record and use your conversations for ads?
Short answer: not without clear consent and transparency. Platforms like Apple, Google, and Amazon require strict privacy controls around voice assistants — and the FTC enforces broad rules against deceptive practices under Section 5 of the FTC Act. The commission said the marketed service would have broken that rule if it worked as advertised and lacked proper consent.
Why local advertisers bought the fantasy — and what sellers sold them
Small businesses were shown a fast path to hyper-targeted leads: one vendor slide suggested targeting neighborhoods where people had mentioned a problem out loud.
You can see the appeal: local advertisers hungry for better ROAS, ad platforms that punish spray-and-pray budgets, and marketers fed a story that voice data is the ultimate signal. In practice, the vendors recycled email lists from data brokers and presented them as a new capability. Their product was a house of cards.
Cox later told Wired that its local marketing team used materials from a third-party vendor, withdrew those materials quickly, and stopped using the product. MindSift and 1010 Digital Works did not provide immediate comment; 1010’s site now redirects to a page with an Albert Einstein quote about intelligence.
What does this mean for ad tech and data brokers?
For anyone who buys marketing services, this should be a wake-up call. Demand proof of data sources, ask how consent was collected, and look for audit trails. Platforms like Google Ads and Meta still rely on signals, but they require compliance with privacy rules; third-party vendors that claim exotic methods should be treated with skepticism.
What you can do right now
Ask for documentation. Insist on vendor contracts that name the exact data partners. Verify opt-in processes and get samples of the datasets being sold to you.
I’ll say it plainly: if a vendor promises voice-level targeting from consumer devices, push back. Check the fine print, call the broker, and if something smells off, walk away.
How will advertisers and regulators adapt when the next-too-good-to-be-true pitch arrives?