Google Ads AI Disclosure Labels: What Service Businesses Need to Know

If you advertise an HVAC company, a plumbing shop, a roofing crew or a dental practice on Google, you picked up a new compliance job this month. Google Ads AI disclosure labels went live in July 2026, and they apply to image and video creative that was generated or edited with artificial intelligence — including the AI images half the contractors we audit are already quietly using.

The good news: for most local service businesses this is a fifteen-minute fix, not a campaign rebuild. The bad news: the responsibility for getting it right sits with you, not with Google, and that was spelled out in the same policy cycle. Here is what changed, what you actually have to do, and what it means for the ads you are running in Phoenix right now.

What actually changed in July 2026

Google updated its advertising policies to require that advertisers disclose when generative AI was used to create or edit an ad's visual assets. The update rolled out gradually through July across Google Ads, Display & Video 360, Campaign Manager 360, Merchant Center and Google Ads Editor, and it introduced a new How this ad was made section inside the My Ad Center panel that consumers can open on any ad they see. The policy detail lives in Google's own documentation at Updates to AI labeling requirements.

The trigger for the change is regulatory, not aesthetic. AI transparency rules in the European Union, India and New York now require that ads containing certain AI-generated or AI-edited assets carry a disclosure so consumers know what they are looking at. Google built one global mechanism rather than three regional ones, which is why an Arizona plumber is now inside the same policy as a Berlin retailer.

Two things are worth being precise about, because a lot of the commentary has blurred them:

• The requirement covers image and video creative — the visual assets. It is not a requirement to disclose that you used ChatGPT to help write a headline.
• The label surfaces in the My Ad Center panel, not stamped across the middle of your ad. Google also now permits advertisers to add their own text or visual labels directly inside the creative if a specific jurisdiction demands it.

Google's own AI tools versus everything else

This is the distinction that decides whether you have work to do.

If you used Google's generative tools inside Google Ads

Asset generation in Performance Max, the image editing tools in the asset library, auto-generated video — when the AI lives inside Google's product, the disclosure is applied automatically. Google knows what it made. You do nothing.

If you used a third-party AI tool

Midjourney, Firefly, Ideogram, Sora, Canva's AI features, an AI headshot generator, a stock photo you had a tool clean up — anything outside Google's walls — and the disclosure obligation is yours. You declare it, using the AI label setting that rolled out through July. Search Engine Journal's coverage lays out the split clearly.

For local service businesses this second bucket is bigger than owners realize. In our audits, the most common offenders are not fancy generative campaigns. They are the small, practical uses: a generated "before and after" mockup because the real job photos were blurry, an AI-upscaled truck photo, a background swap on a team headshot, a stock model dropped into a fake kitchen. All of those are edits, and all of them count.

The other half of the story: who is responsible

The labeling policy did not arrive alone. On July 1, 2026, Google also updated its Google Ads Terms of Service to explicitly authorize its automated systems to format, select and generate ads, targets and destinations on the advertiser's behalf. The same revision reinforces that advertisers remain responsible for reviewing, approving, editing or removing campaigns and assets — including the ones Google's own tools generated. Search Engine Land covered the terms update ahead of the rollout.

Read those two changes together and the picture is blunt. Google gets broader latitude to automate. You keep the liability. If an auto-generated asset shows a plumber standing in front of a house your company never worked on, or a headline promises same-day service you cannot deliver, that is your problem — legally and reputationally.

This is the same accountability gap we flagged when Google pushed more inventory into automated formats. If you are still deciding how much control to hand over, our breakdown of Performance Max versus Search campaigns for home services is the practical version of that decision.

What to do this week

1. Inventory your creative

Open your asset library and go image by image. For each one, answer a single question: did a human photograph this exactly as it appears, or did software generate or alter any part of it? Upscaling, background removal, generative fill, object removal, and AI "enhancement" all count as edits. Flag anything you are unsure about — over-disclosing costs you nothing.

2. Apply the label setting

For every flagged third-party asset, turn on the AI label. If your creative runs in the EU, India or New York, check whether an in-creative visual label is also required rather than the panel disclosure alone.

3. Start a creative log

One spreadsheet: asset filename, date added, tool used, whether AI touched it, who approved it. This takes ten minutes a month and it is the only thing that will save you if a disclosure question ever comes back at you six months from now. Agencies handling multiple locations should keep this per account, not per client.

4. Audit whether you need AI creative at all

Here is the uncomfortable question. Most service business ads perform better with real photography — your actual trucks, your actual crew, your actual finished driveway — than with generated imagery that looks like every other contractor's. If your only reason for using AI images is that nobody has taken good job photos in three years, the policy change is a nudge to fix the root cause. Give a tech a phone and a two-line checklist and you will have better creative in a fortnight.

Will the label hurt performance?

There is no meaningful data yet, and anyone quoting a number is guessing. What we can say from the structure of the rollout: the disclosure sits inside My Ad Center, a panel the overwhelming majority of consumers never open. The realistic risk is not click-through rate. It is trust in categories where authenticity is the whole product.

Home services is exactly that kind of category. A homeowner letting a stranger into their house at 9pm is buying trust, not pixels. If a prospect does dig in and finds that your "our recent work" image was generated, the damage is not a lost click — it is a lost booking. Weigh it accordingly.

Where this is heading

The direction of travel is obvious: more automation from Google, more disclosure obligation on advertisers, more of your account decided by systems you do not directly control. The counterweight is measurement. If you cannot see which assets and which campaigns produce booked jobs, you cannot tell whether the automation is helping. That starts with conversion tracking that actually reflects real jobs, and it ends with landing pages built to convert the traffic you are paying for.

If you would rather someone else did the creative inventory, the label setup and the account review, that is what our free Google Ads audit covers. We will tell you which assets need disclosing, where automation is quietly spending your budget, and what we would change first. No obligation, and you keep the findings either way. You can also see how we handle ongoing account management on our Google Ads management page.