New York's AI Ad Disclosure Law And Three More Ecommerce Ad Updates You Need To Know (July 2026)

New York's AI Ad Disclosure Law And Three More Ecommerce Ad Updates You Need To Know (July 2026)

There's a new law on the books that can fine you $5,000 for ads you might be running right now.

The New Law That Can Fine You $5,000 For AI Ads

New York has become the first US state to legally require advertisers to disclose when they use AI-generated or synthetic performers in their ads.

It took effect on the 9th of June, so it's already live. The penalty is up to $5,000 per violation. This was reported by the Associated Press.

Let me be clear about what it covers, so you don't panic if it doesn't apply to you.

This is about synthetic performers: AI avatars, AI actors, AI-generated spokespeople, or synthetic voiceovers that sound like a real human talking about your product.

If you're running an ad where an AI-generated person holds up your product and tells the camera how great it is, and that ad shows to people in New York, the law now says you need to disclose that it's AI.

Here's why this matters even if you're nowhere near New York:

A huge number of ecommerce brands have leaned on AI-generated UGC over the last year.

It's cheap, it's fast, and you can spin up a dozen talking-head ads in an afternoon without hiring a single creator. And loads of you run national campaigns on Meta and TikTok, which means your ads almost certainly show to people in New York whether you targeted them specifically or not. So this applies to far more of you than you'd think.

The experts quoted in that coverage all expect other states to follow New York's lead. That's almost always how regulation spreads in the US: one big state goes first, then a wave of others copies the template over the next year or two.

There's also a really interesting stat buried in this story. Consumer enthusiasm for AI-generated ad content has fallen off a cliff. Back in 2023, around 60% of people were positive about AI in ads. By 2025, that had dropped to just 26%.

So here's what I'd actually do this week.

First, audit your live ad creative for anything using an AI avatar, an AI actor, or a synthetic voiceover. Second, for those ads, add a clear disclosure that the content is AI-generated. Don't wait for a $5,000 fine to force your hand. And third, honestly, given where consumer sentiment is heading, I'd think hard about how much you're leaning on fully AI-generated people in the first place. Real customers, real creators, and real founders talking about your product still tend to convert better anyway.

Prime Day Hit A Record $26.4B, But Read The Small Print

Amazon's Prime Day event this year generated $26.4 billion in total US ecommerce sales, and day one alone reached a record $8.3 billion. This came from Digital Commerce 360, and on the face of it those are enormous, record-breaking numbers.

But when you actually read the small print, the story becomes a lot more complicated. Even though it was a record in absolute terms, the year-on-year growth was the slowest on record. And average basket sizes actually shrank. So people were shopping, but spending more cautiously and buying less per order.

Here's the part that's genuinely useful for you. An affiliate platform called Levanta tracked the same 390 brands across two Prime Day events. From one Prime Day to the next, Amazon ad ROAS actually fell by 12.5%. But affiliate ROAS for those same brands climbed by 5.3%.

Here's why that matters.

During these big peak events, everybody piles into Amazon ads at the same time. When everyone's bidding on the same inventory on the same day, the auction becomes more expensive and your return drops. That's exactly what that 12.5% fall is showing you. Meanwhile affiliate marketing, which is people recommending your product through content and links, became more efficient.

So the lesson here isn't that Amazon ads are bad. It's that during peak events, leaning on one single channel is where you're squeezed.

Whenever there's a big shopping event, whether it's Prime Day, Black Friday, or a seasonal push, don't just dump all your budget into the most obvious, most crowded channel. Diversify where the demand comes from. Lean on affiliates, lean on content, lean on the channels where you're not fighting every other advertiser on the planet for the same click.

And watch your efficiency, your ROAS and your margins, not just the big top-line revenue number. A record sales day that quietly costs you profit on every order isn't the win it looks like on the headline.

The $10M Ad Platform That Just Opened To Everyone

There's a big ad platform called AppLovin that, up until now, you basically couldn't use unless you were a massive advertiser. It was invite-only, with a minimum spend of $10 million or more just to walk through the door.

Well, that's just changed. AppLovin has removed the gate and opened up self-serve access to everyone. What caught my eye is that early incrementality tests show it outperforming other channels for some advertisers.

AppLovin is a big player in the mobile and app advertising world, and the reason this matters for your store is all about diversification. Right now, the vast majority of you run everything on Google and Meta. I understand why; that's where the volume is. But relying on just two platforms means you're completely at the mercy of their auctions, their price rises, and their reliability.

And here's the bigger picture. This AppLovin move landed in the same month that Walmart and Fox both made big moves to consolidate connected TV ad inventory. So there's clearly a wider land-grab going on for advertising budget that sits outside of just Google and Meta.

So should you test AppLovin? Here's my honest take.

I manage over $12 million a year in ad spend, and one thing I've learned is that the advertisers who win long-term are the ones always testing the next channel before they're forced to. But you test it properly. You don't move your budget over on a hunch.

The key word in that story was incrementality: measuring whether the new channel is driving extra, genuinely new sales, or just taking credit for sales you'd have made anyway. Carve off a test budget, set up your conversion tracking properly, and measure the incremental lift. If it's genuinely adding new customers on top of Google and Meta, brilliant, scale it up. If not, you've learned that cheaply and you move on.

Google's Smart Bidding Exploration Just Went Global

Last one, and this is a Google update that can genuinely help you find new customers you're currently missing. Google has announced that Smart Bidding Exploration is now available globally, in all languages. This came straight from Ginny Marvin, Google's Ads Liaison.

Let me explain what it does, because the name is a bit of a mouthful. Normally, when Google's Smart Bidding runs your campaign, it bids on the searches it's already confident will convert for you. Smart Bidding Exploration lets it go a step further and reach new, valuable searches that are currently out of its reach, without you having to change your core targeting. Think of it as Google saying, "there are searches out here you're not showing up for, but I reckon they'll convert for you, let me go test them."

It's now available for Performance Max campaigns that don't have a product feed, and for Search campaigns. There's also a new beta opening up for Shopping ads, both on Performance Max with a feed and on Standard Shopping campaigns. So this touches basically all the campaign types you're most likely running.

One of the hardest parts of scaling is finding genuinely new customers, rather than just paying more to reach the same pool of people over and over. That's exactly the problem this is designed to help with. My honest take is that this is worth turning on and testing, but you watch it like a hawk. Enable it on a campaign that's already performing well and has a bit of headroom in the budget, then keep a close eye on your search terms report and your efficiency over the following few weeks. If it's bringing in new, relevant searches that convert profitably, fantastic, that's new growth. If you see it wandering off into irrelevant searches, you rein it back in. Used carefully, it's a genuinely useful way to find the new customers your current targeting is quietly missing.

Conclusion

Four stories worth your attention this month. New York became the first US state to require advertisers to disclose AI-generated or synthetic performers in their ads, with fines of up to $5,000 per violation and other states expected to follow, so audit your creative and add clear disclosures now. Amazon's Prime Day set a record $26.4 billion, but the slowest-ever growth and a 12.5% drop in Amazon ad ROAS (against a 5.3% rise in affiliate ROAS) prove why you should never pour all your budget into one crowded channel during peak events. AppLovin scrapped its $10 million minimum and opened self-serve to everyone, a reminder to test new channels on true incrementality rather than betting the farm. And Google's Smart Bidding Exploration went global across Search, Performance Max, and a new Standard Shopping beta, a handy way to reach profitable new searches as long as you monitor your search terms closely.