If a vendor is pitching “AI targeting” and the demo sounds like it involves your customers’ phones listening to them, stop right there. This week the FTC treated an “Active Listening” marketing pitch like deception, not innovation.
What the FTC alleged (in plain English)
On May 21, 2026, the FTC announced proposed settlements involving Cox Media Group, MindSift, and Wisconsin-based 1010 Digital Works. The FTC alleged the companies claimed they could target localized ads based on conversations captured from consumers’ smart devices, and that consumers had opted in. The FTC also alleged that was not how the service worked.
If you want the primary sources, start here:
- FTC press release (May 21, 2026): “Active Listening” case summary
- FTC complaint (CMG / Cox Media Group)
- FTC complaint (MindSift)
- FTC complaint (1010 Digital Works)
“Creepy? Sure. Great for marketing? Definitely.”
That line appears in the FTC’s CMG complaint. No, we are not making it up.
The real takeaway is boring
Smart speakers are the hook. The real issue is marketing claims. If you sell something using a story that makes the buyer feel clever, and the story is not true, you are buying risk. You might also be buying a reputation problem you cannot undo with a nice apology post.
Steal this checklist before you buy any “AI targeting” service
Ask these questions on a call. Then ask for them in writing. If the answers turn into vibes and hand-waving, that is your answer.
1) What data is actually used?
- “Voice data” is a claim. Ask what the inputs really are.
- Ask whether the output is a list, an audience file, a pixel, a report, or just a dashboard.
- Ask for a sample deliverable. A real one, with private details removed.
2) Where does consent happen?
Do not accept “it’s in the terms” as an answer. If a service relies on opt-in consent, the vendor should be able to show you the opt-in.
- Show me the exact screen or flow where the user agrees.
- Show me what the user is told, not what your lawyer wishes they were told.
- Show me how opt-out works, and who handles the request.
3) What are you paying for?
Some pitches are just list buying with a costume. That does not automatically make it useless, but it changes what you should expect and what you should claim.
- Are you buying third-party lists? If yes, name the broker.
- What is the markup? What is the minimum buy? What is the refund policy?
- If the list is wrong or stale, what happens next?
4) What will you let us say publicly?
Here is the test. If you cannot describe how the service works without sounding creepy or dishonest, do not put it in your marketing. And do not let a vendor write that copy for you.
If you are the vendor, fix your pitch deck
If you sell advertising or marketing services, stop using sci-fi language to make ordinary targeting sound magical. You will close a few deals. You will also build a file on yourself.
- Do not imply you are listening to consumers’ conversations.
- Do not claim “opt-in” unless you can point to the opt-in.
- Do not pretend a nationwide list is “local” targeting.
- Do not use “AI” as a substitute for being specific.
Want a less risky way to reach ready buyers?
Here is the unsexy play that still works: tighten your proof, tighten your offer, tighten your lead response, and run clean retargeting on people who actually visited your site.
Start with a case study format that reads like real work, then build a lead response system so you stop leaking the leads you already paid for.
If reviews are part of your trust engine, read our breakdown of the FTC’s fake reviews rule too. Different topic, same lesson: do the boring parts right.
Need help cleaning up marketing claims and building a safer growth plan?
We help businesses and nonprofits get specific, build proof, and run marketing that does not require a legal-adjacent bedtime story.