You've spent years managing what Google says about your centers. The FTC just started caring what the AIs say.

Buried in the FTC's July 1 release stack — under the settlements and the Made-in-USA warning letters — is a nine-page proposed policy statement that I think is the most consequential document of the month for anyone doing what you do. I read the PDF, not the coverage. Its core claim, in the Commission's own words:

That sentence is the federal government asserting, for the first time in a formal policy document, that the accuracy of what an AI says is consumer-protection territory — Section 5, the same law behind every deceptive-advertising case ever brought. The statement's target is AI companies steering their models ("steering an AI system in this manner may deceive consumers in violation of Section 5 of the FTC Act"), and it squares up against state AI laws directly — Colorado's Act is named, with the FTC's answer that state law is "impliedly preempted to the extent it conflicts with a federal regulatory scheme."

Here's my read for your world: when patients ask ChatGPT or Google's AI where to get an MRI, what those systems say about your imaging centers — prices, capabilities, locations, credentials — is drifting from "unfortunate when wrong" toward "legally salient when wrong." A regulator has now put accuracy-of-AI-outputs on paper as a duty someone owes consumers. That is the strongest tailwind the get-your-facts-right-where-AIs-read-them argument has ever had.

Kept honest: this is a PROPOSED statement (2-0 vote), aimed at AI companies steering outputs for undisclosed ideological ends — it does not regulate marketers, and it can change after comments. The direction is the story: AI-output accuracy now has a docket number and a deadline, and directions like this rarely reverse.

Sources (primary, verified today): FTC, Proposed Policy Statement Concerning the Suppression of Accuracy in Artificial Intelligence Systems (July 1, 2026, 9 pp.) — all quotes verbatim from the PDF · FTC press release — comment deadline July 31, 2026; Commission vote 2-0; Federal Register publication pending; directed by a December executive order. Read the statement, not the coverage — it is nine pages and it names names (Colorado's AI Act).

You've spent years managing what Google says about your business. As of July 1, the federal government formally cares what the AIs say.

Buried in the FTC's release stack is a nine-page proposed policy statement on AI accuracy. I read the PDF, not the coverage. The load-bearing sentence, verbatim:

"Consumers have a reasonable expectation that AI systems aim to give truthful and accurate outputs."

That's the first formal federal policy document putting AI-output accuracy under Section 5 — the same law behind every deceptive-advertising case ever brought. It targets AI companies steering their models, and it takes direct aim at state AI laws (Colorado's is named) with a preemption argument.

Why I think this matters if you market healthcare services: when a patient asks an AI where to get an MRI, what it says about prices, locations, and credentials is drifting from "unfortunate when wrong" to "legally salient when wrong." The get-your-facts-right-where-AIs-read-them work just acquired a federal tailwind.

Three moves:
1. Comments close July 31. If you've watched AI systems get client facts wrong, a regulator just formally asked to hear about it.
2. Baseline now — dated captures of what the major AIs say about each client. Baselines turn anecdotes into evidence.
3. Mind your own claims: promises about an AI's accuracy are Section 5 material too. Say what you can prove.

Proposed, not final — 2-0 vote, and it can change after comments. But accuracy-of-AI-outputs now has a docket and a deadline. Directions like this rarely reverse.

FIRST SOURCE · one verified original-source finding, composed for one reader · this edition: radiology marketing