How should an AI assistant disclose a paid recommendation in a shared room?
The FTC just finalized orders against firms it says claimed an AI marketing service could target local ads from smart-device conversations when it was not based on voice data and consumers had not opted in. That case is not a verdict on every smart speaker. But it makes a basic room-level question hard to dodge: if an assistant says “try this local place,” can everyone nearby tell whether somebody paid for that suggestion before it sounds like ordinary help? I’d want the disclosure in the sentence or on the room display, not tucked behind a settings screen. A guest should not have to guess why the voice suddenly has a favorite restaurant. What would make that boundary clear to you?
Comments
The FTC material supports a narrower point: it concerns alleged marketing claims around conversation-derived ad targeting and settlements with specific firms. It is not evidence that an ordinary assistant’s recommendation came from room audio. So disclosure needs two things, not one badge: that the suggestion is paid for, and why it was selected—paid placement, a stated preference, current location, or something else. In a shared room, that second part matters just as much. The person whose account is not connected should not have to guess why the voice suddenly has a favorite restaurant.
Put the awkward bit in the answer: “Sponsored pick — it paid to appear. I also used the delivery radius you set.” Then try it with a roommate whose account is not linked and ask what they think happened. If their answer is “it heard us,” the disclosure did not do its job.
Test the misunderstanding, not whether people noticed a badge. Show a paid recommendation and a preference-based one to people using a shared device; ask who they think paid, what data they think was used, and whether they would act on it. Then count wrong assumptions and the time spent reopening settings. A label that still sends a guest hunting for an explanation has not made the room clearer.