A recent consumer survey from the National Advertising Division is pretty damning: 78% of people feel like they’ve been duped by AI-generated ads that weren’t clearly labeled (NAD 2025 AI Transparency Report). That level of distrust is exactly what Governor Newsom’s new disclosure law is meant to tackle. The mandate means we have to clearly identify AI-generated parts of our ads, and that completely changes how we approach creative and measure campaign performance. The big question for everyone in the trenches is what this does to our numbers.
Key Takeaways
- In California, you now have to put clear, obvious disclosure labels on all AI-generated ad content, which is already disrupting creative production workflows.
- Early data from Q1 2026 is showing a 15% average drop in click-through rates for ads that disclose AI use versus those that don’t, forcing us to rethink strategy.
- To stay compliant, you need to run detailed internal audits on your content process to find and tag every AI-assisted element. It’s a new legal minefield.
- The law is going to force more spending on AI detection tools for both advertisers and regulators, bumping up operational costs for many firms.
- Building consumer trust with authentic content and being transparent about AI is the only path forward. Trying to hide it brings huge reputational and legal risks.
Data Point 1: 35% Increase in Compliance Costs for California-Based Advertisers
Since the law went into effect on January 1, 2026, the immediate hit has been to the budget. A preliminary report from the California Advertising Regulatory Board (CARB 2026 Compliance Impact Report) shows that companies saw their compliance-related spending shoot up by an average of 35% in just the first quarter of 2026. This isn’t pocket change. We’re talking about the cost of new software, extra legal reviews, and more people on staff just to vet ad content. That means licensing AI detection tools like AI Authenticity Scanner and even hiring dedicated compliance officers. This is a massive overhaul of the ad production pipeline for a lot of shops. For smaller agencies, a cost burden like that could easily stifle innovation if they don’t manage it carefully.
Data Point 2: 15% Dip in Click-Through Rates for Disclosed AI Ads
For performance marketers, the most glaring metric is the immediate hit to user engagement. Early data out of Q1 2026, pulled together by the Digital Advertising Alliance (DAA Q1 2026 Report), points to a 15% average decrease in click-through rates (CTRs) for ads that are properly labeled as having AI content, compared to similar ads without the disclosure. That’s a real consumer reaction, not just a blip. My take? People are still suspicious. They see an “AI-generated” tag and assume the content is less authentic or trustworthy, even if the ad itself is good. Campaigns that lean heavily on AI for their creative assets have to pivot, period. This means we have to get back to basics, focusing on the core value proposition and the human element in the copy, and probably start A/B testing different disclosure phrasing and placements to see what softens the blow.
Data Point 3: 25% Increase in “Authenticity” Keywords in Ad Copy
Advertisers are already scrambling to adjust their messaging in response to the disclosure rules and that CTR dip. A linguistic analysis from the Ad Research Institute (Ad Research Institute 2026 Trends) of thousands of California ad campaigns from Q1 2026 found a 25% jump in the use of keywords like “authenticity,” “real,” “human-crafted,” and “genuine” in ad copy. It’s an obvious attempt to get ahead of consumer skepticism. It’s also a smart psychological play, if you’re forced to admit you used a robot, you shout even louder about the human parts. We’re seeing brands talk up their artisanal qualities, use more (verified) real user testimonials, and show behind-the-scenes content of their actual human teams. It’s a tacit admission that while AI is efficient, that human connection is still what builds trust and gets conversions.
Data Point 4: 10% Rise in Consumer Reports of Undisclosed AI Ads
Even with the new regulations, enforcement is tough, and it looks like consumers are stepping up to be the watchdogs. The California Department of Consumer Affairs (CDCA 2026 AI Ad Complaints) saw a 10% increase in consumer complaints about ads with potentially undisclosed AI content just between January and March 2026. That 10% might not sound like a huge number, but it shows a real change in public awareness and a new willingness from people to police ads themselves. This is the real test for brands. One misstep can lead to public backlash, regulatory fines, and lasting damage to your reputation. The days of quietly weaving in AI are gone. Brands that try to get away with it are going to find themselves in the crosshairs of a newly empowered and very watchful public. The signal is clear: transparency is now a consumer expectation, not just a line item in a legal doc.
Challenging the Conventional Wisdom: Disclosure Isn’t a Performance Killer, It’s a Trust Builder
Many marketers I talk to are convinced that AI disclosure is a performance killer. And sure, the 15% CTR dip seems to prove their point. But I think that’s a shortsighted take. While the initial numbers show a dip, I believe it’s a short-term market correction. The knee-jerk reaction to any new rule is always focused on the immediate, negative impact. But what these Q1 2026 figures really show is a market adjusting to a new standard of honesty. People aren’t against AI. They’re against being deceived. This law, despite the upfront performance headaches, is forcing brands to build a much stronger foundation of trust. Think it through: once consumers get used to seeing these disclosures and brands figure out how to integrate them without being awkward, the ads *without* a disclosure will start to seem suspicious. The brands that lean into transparency right now, even if it means taking a temporary hit to CTR, are the ones that will build deeper, more loyal customer relationships for the long haul. It’s a strategic investment in ethical marketing. Calling this a “performance killer” totally misses the critical long-term value of consumer trust as AI becomes a bigger part of our world.
Newsom’s law is a reset for AI advertising, demanding transparency and accountability from everyone. The initial data shows costs are up and engagement is down, but these are growing pains. In the end, this pushes the industry toward a more ethical future, one where consumers might actually start to trust AI-generated content.
So what exactly has to be disclosed as AI under the new CA law?
The law covers any significant part of an ad, visuals, audio, or text, that was substantially generated or altered by artificial intelligence. This means things like deepfakes, AI-synthesized voices, or AI-written copy that forms a core part of the ad’s message need a disclosure.
How are we supposed to implement these disclosures correctly?
You have to use labels that are clear and easy to see, like “AI-Generated Content” or “Partially AI-Assisted,” and place them somewhere prominent in the ad. For video or audio ads, this usually means a quick disclaimer at the start or end, or even a persistent text overlay on the screen.
Are there any exceptions for small uses of AI?
The law is focused on “substantial” AI generation. So minor uses, like using an AI-powered grammar checker or basic photo touch-ups that don’t fundamentally change the image’s meaning, probably don’t need a disclosure. But the line can be blurry, so it’s always best to check with a lawyer.
What are the penalties if you don’t comply?
Failing to comply can get expensive. You could be looking at big civil fines for each violation, cease-and-desist orders, and even being forced to run corrective ad campaigns. The exact penalties are decided by regulators like the California Attorney General’s office.
How does this law change A/B testing with AI content?
It adds a new layer to your testing strategy. Now, you have to factor the disclosure itself into your A/B tests. You’ll be comparing disclosed AI ads against your non-AI controls, and you might even test different disclosure methods against each other. It makes optimization more complex, but also more honest.
