A question addressed to platforms, landing on advertisers
On September 24, 2026, the US Federal Trade Commission voted 2–0 to publish an advance notice of proposed rulemaking. It asks whether to extend its 2024 rule on impersonation of government and businesses to cover platforms whose ad-optimization tools help impersonation scams reach people. The options on the table include vetting advertisers, monitoring ads, investigating suspected scam ads and removing confirmed ones. The FTC's own data explains the urgency. Consumers reported nearly $3.5 billion in impersonation losses in 2025. Nearly 30% of people who lost money to scammers said the first contact came through social media, and those losses alone came to $2.1 billion. There is no rule text yet, as Mayer Brown's analysis points out. Comments are due 60 days after the notice appears in the Federal Register.
The notice is aimed at scammers and at the platforms that serve them. Legitimate advertisers are not its subject. But the question underneath it does not stay neatly in its lane: when a system writes, combines, targets and optimizes an ad, who checked that what it says matches where it leads?
Questions worth holding open
The useful response to a question like that is not to rush to an answer. It is to ask the next question, and then the one after. These are the ones we keep returning to. We have not answered them, and we do not think they answer easily.
If the optimization system chose the headline, the image and the audience, and you chose the budget, which of you made the claim? Your name is on the ad. The combination that ran may be one you have never seen. Is authorship about who wrote the parts, or about who released the whole?
Can an ad be misleading when every element in it is true? A real partner logo, an accurate discount and the word "official" can each be defensible on their own. Put together by a system that only knows they performed well, they may imply an endorsement nobody gave. Where does the meaning of an ad live: in its parts, or in what a reader concludes from them?
You approved the ad. Did you approve the version that ran? Dynamic creative, automatically generated text and placement-specific crops mean the approved asset and the delivered impression can differ. If you cannot say which version a given customer saw, what exactly did your approval cover?
If platforms start examining what sits behind ads to protect themselves, whose standard of "matching" applies? A platform reviewing destinations at scale will need a definition of consistency. It may be stricter than yours, looser than yours, or simply different. Are you prepared to argue that your page supports your ad, using a standard you have never written down?
What does knowing actually cost, compared with assuming? Most teams assume alignment because checking feels expensive. That calculation was made when nobody else was checking. Does it still hold once someone else might be?
And if the answer to all of this is "the platform should have caught it", what did you hand over along with the budget? Automation was sold as a way to delegate effort. It is less clear that it delegates judgment, and less clear still that it delegates blame. A media buyer who sets a target CPA and lets the system find the words has made a choice about what to optimize for. Whether that choice includes an obligation to look at what the system then said, in the advertiser's name, is a question nobody has had to answer explicitly until now. Perhaps it never needed answering. Or perhaps it was always the real question, and volume simply hid it.
It is tempting to resolve each of these with a policy line and move on. We would rather leave them uncomfortable a while longer, because the discomfort is accurate. The tools changed faster than the definitions of responsibility, and nobody has caught up yet: not platforms, not regulators, and not most advertisers.
Examined alignment: what the questions mean in practice
The FTC is unlikely to write a rule that governs your Black Friday ads. The more probable effect on legitimate advertisers is indirect. Platforms facing regulatory pressure on their optimization layer have an obvious incentive to tighten automated review before any rule arrives. That would likely mean more scrutiny of advertiser identity, and more checks that compare an ad's claims with its destination. Google already disapproves ads under its misrepresentation policy when the destination does not support what the ad implies. A wider net, cast by automated systems, will catch some honest ads whose pages simply fail to prove what the ad says.
The ads most exposed share a few traits:
- They use third-party brand names, partner logos or certification marks.
- They use government-adjacent language: "official", "approved", "rebate", "eligible".
- They pair a specific offer with a page that states it more vaguely, or not at all.
- They run as automatically assembled variants that nobody has reviewed as a combination.
None of these is wrongdoing. Each is a gap between what the ad suggests and what the page demonstrates. That gap used to cost conversions. It may soon cost delivery too.
This is where the questions above become operational. You cannot answer "which version ran?" without a record of what ran and where it pointed. You cannot argue that your page supports your ad without a consistent definition of support. And you cannot weigh the cost of knowing against assuming until you have measured how often your assumptions are wrong. In our experience, teams that measure are usually surprised. The signs of ad-to-page misalignment are rarely dramatic. They are a vague claim here, a missing logo there, a promise moved below the fold.
AdAlign was built to replace that assumption with evidence. It scores each ad against its landing page across four dimensions, Visual Match, Message Match, Above the Fold Continuity and Tone Alignment, and lists the specific claims the page does not substantiate. Each finding is a sentence, so you can act on it; each pair gets one score out of 10, so you can track it over time. That history is the part creative governance has usually lacked: a dated record that each ad and its destination said the same thing, as opposed to a belief that they probably did. If the underlying concept is new to you, start with what ad-to-page congruence measures.
There is an obvious objection. If the FTC rule never happens, was the effort wasted? We do not think so. The same mismatches a platform reviewer would flag are the ones that make a visitor hesitate. Better message match reduces regulatory exposure and improves conversion with the same fix. And as ads keep changing weekly while pages change quarterly, alignment drift means the evidence has to be refreshed, not filed once.
The court ruling we wrote about in September made the same point from another direction: what your landing page links to now counts as part of your ad. Accountability is spreading outward from the ad, toward the page and toward the systems that assemble both.
For the practical side of this quarter, our pre-season holiday promo drift audit shows how to check offer claims against pages step by step. For the case where the page drops out of the purchase entirely, see our analysis of Google's AI Mode native checkout.
Answer the question before someone else asks it
Before a platform's reviewer reads your ad against your page, read it yourself. Run a free audit on one ad: a score out of 10 across four dimensions, and a list of the mismatches the page needs to fix. No call, no card.
Frequently asked questions
What is the FTC's September 2026 notice on impersonation scams? On September 24, 2026, the FTC issued an advance notice of proposed rulemaking asking whether to update its Impersonation Rule to address platforms whose ad-optimization tools help impersonation scams spread. It is an inquiry only; no rule text exists yet.
Does the FTC inquiry affect legitimate advertisers? Not directly. The more likely effect is indirect: platforms may tighten automated review of advertisers and destinations, which can flag honest ads whose landing pages do not clearly support the ad's claims.
What is a misrepresentation disapproval in Google Ads? Google can disapprove or restrict ads under its misrepresentation policy when an ad or its destination is likely to mislead, for example when the landing page does not support what the ad implies about an offer, affiliation or endorsement.
How do I make sure my AI-generated ads match my landing page? Keep a record of every ad variant that runs and the destination it points to, score each pair on visual match, message match, above-the-fold continuity and tone, and fix any claim the page does not substantiate. Re-check whenever the ad or the page changes.