The silence around AI generated ads is ending.
In July 2026 Google announced a new AI disclosure for ads across Search, YouTube and Discover. A "How this ad was made" panel shows whether an ad was created or edited with AI. Google adds the disclosure automatically when its own generative AI tools made the ad, and advertisers who use other AI tools get a new control to indicate it.
The panel sits in My Ad Center, which anyone can open from the three dot menu on an ad. For creative made outside Google's tools, TechCrunch reported that Google will not run its own check. It takes the advertiser's word.
That makes it sound like a checkbox, one more field in the setup flow that someone on your team clicks through.
It is bigger than that. You now need to know, and record, how every ad was made. And real photography of your product and your people has become a competitive advantage where it used to be a cost you tried to avoid.
Until now, whether your ad imagery was photographed or generated was something only you knew. Today Google asks you to declare it, and anything a business declares is something its customers can look up.
What this changes for advertisers
Three practical things, starting now.
You need to know where your creative came from. Many businesses do not. An agency made some of it. A freelancer made some. Somebody used a generator for a background. A platform tool suggested a few variations. When the disclosure control asks about AI, "we are not sure" will not do.
Your creative process now has a compliance step. Whoever builds your ads needs to record which tools touched each asset, because the person setting that control has to get it right.
Your brand has a new position available. "We shot this for real" is now a distinction Google itself asks advertisers to draw.
When every advertiser is asked whether a machine made the ad, the ones who can honestly say no have something their competitors cannot buy back with a prompt.
Why real just became an advantage
Generated imagery is cheap and fast. So everyone uses it, and it no longer sets anyone apart.
When anyone can produce a glossy, perfectly lit lifestyle shot in seconds, the glossy shot stops telling the buyer that this business invested in showing the product properly.
A real photograph tells them several things at once. This product exists and looks like this. These are the people who make it or use it. This is the place. Somebody turned up.
Buyers want that proof more than ever, because they can no longer be sure what in their feed is real.
Think about how a buyer decides. They see the ad, then they check. They look at your reviews, your Google Business Profile, your Instagram, the photos other customers posted. If the ad matches what they find, trust goes up. If the ad shows a sleek studio kitchen and your real photos show a busy suburban workshop, the gap does the damage, even when nothing in the ad was false.
Real creative closes that gap by default, because the ad and everything the buyer finds afterwards show the same business. No single perfect image persuades like that match does.
We have written before about authenticity as a creative style. Disclosure turns it into a recorded difference in how your ads are made.
There is a downside risk to weigh as well. A generated shot that flatters a $90 lamp sets up a gap between the ad and the parcel that arrives. We covered that trap in why fake product photos can torch your brand. The risk was there before disclosure. Now it is harder to ignore.
The creative audit to run this month
Before you think about positioning, find out what you are running.
- List every active ad asset across Search, YouTube and Discover, plus anything you plan to launch.
- Record its origin. Photographed, filmed, designed by a person, or created or altered with AI tools. Note which tools.
- Check Google's current wording for where the line sits between creating and editing. The announcement does not define it. Apply whatever rule you settle on consistently, and when unsure, ask Google Ads support.
- Brief your agency and freelancers in writing. They must tell you whenever AI tools are used on anything they deliver.
- Store the origin in your asset library so the information survives staff changes.
How to set up a creative origin log
The audit is a one off. The log keeps you right every month after it, and a shared spreadsheet is enough if everyone uses it.
- Give every asset an ID. Use the same name in the file, the log and the ad platform, so you can match them in seconds.
- Add five columns. Who made it, the date, how it was captured, which tools edited it, and whether any element was generated.
- Record at the point of creation. The person who makes or edits the asset fills in the row before handing it over. Reconstructing history later is where errors creep in.
- Link to the raw file. For real photos and video, keep the original capture. It is your evidence if anyone ever asks.
- Make it a launch gate. No asset goes live without a completed row. The person setting up the campaign reads the log before touching the disclosure control.
Add one column for rights as well: who owns the image and where you are licensed to use it.
Shoot real where it counts
Nobody is asking you to swear off every tool. Set a clear rule for what must be real.
The product. Always. What people see in the ad should be exactly what arrives. Texture, colour, scale and finish, photographed as they are.
The people. Your team, your founder, your customers. Faces carry trust, and buyers can feel the contradiction when a real business advertises with a synthetic face.
The results. Before and after shots and finished jobs. If it is presented as evidence, it has to be evidence.
The place. Your shopfront, your workshop, the job site, the street outside. Local businesses in particular win by being recognisably from somewhere.
AI still has a job behind the scenes. Mood boards before a shoot. Shot lists. Rough copy options for a human to rewrite. Keep it away from anything presented as proof.
One test settles most arguments. Would a customer feel misled if they found out how the image was made? Then it has to be real. If they would not care, as with a generated abstract background behind a real product photo, you have more room, as long as you log it and disclose it.
Making real affordable
The objection is always cost. With planning, it shrinks.
Batch it. One well planned shoot day can supply product, team, place and process imagery for months of ads. We break down how in one video shoot, 90 days of content.
Shoot for every placement. Brief the photographer to capture wide, tight, vertical and square compositions of each setup, so one image is not cropped badly into every format.
Capture video and stills together. Short real clips cost little extra when the setup already exists, and you will use them everywhere.
Keep a phone habit. Between shoots, capture real moments in the business every week. A slightly rough photo of Tuesday's delivery still beats a generated one.
Own the rights. Make sure your agreement with the photographer covers advertising use across every platform you run, for as long as you need.
Build a library you can search. Tag every image by product, person, place and format so the team can find the right shot in a minute. When real imagery is slow to find, people reach for a generator.
A worked example: one shoot, a quarter of ads
This is a hypothetical, to show how the maths changes.
Say you run a Perth homewares store running YouTube and Discover ads, and most of your current creative is generated lifestyle scenes. You book a single shoot day in your own store. The shot list covers your ten best sellers in natural light, two staff members styling a shelf, a customer unboxing at the counter, and the street frontage in the morning.
From that one day you have stills in four ratios for every product, short vertical clips for YouTube, and a set of "real store, real people" images for Discover. Every asset goes into the log as photographed, with no generated elements. Disclosure takes seconds, and your ads show what a customer will find when they walk in.
Compare that to the generated approach. It is faster on day one. After that, every asset needs checking against the policy, every product shot risks drifting from the real thing, and none of it tells the buyer anything true about your store.
Generated material still earns a place in planning the shoot or mocking up a concept before you commit. In this scenario it stays in the planning folder, and the ads that carry the budget are the real ones.
What disclosure does not mean
AI assisted ads are still allowed, and a business using AI tools is doing nothing wrong by using them. In the same announcement Google says it keeps prohibiting misleading and deceptive ads, whether AI made them or not.
What has changed is that honesty about process is now expected. Use the tools openly, disclose them properly and you are fine. The businesses that get into trouble will be the careless ones who never tracked how their ads were made.
Nor should you panic about every retouch. Removing dust from a product shot or correcting the colour balance of a real photo is a long way from inventing a scene. Because Google has not published a line between the two, write your own internal rule so nobody is guessing case by case.
The move
Businesses that treat disclosure as a nuisance will set the control and carry on. Treat it as a signal and you will build a library of real creative your competitors cannot match.
If you want that library planned around your ads and shot for every format you run, our photo and video team does that work.
Pull up your highest spending ads today and write down how each one was made. If you cannot answer with confidence, that is your first job. If most were generated, your next shoot is overdue.
Frequently asked questions
Do I have to disclose AI generated images in Google Ads?
In July 2026 Google announced a "How this ad was made" panel for ads on Search, YouTube and Discover that shows whether an ad was created or edited with AI. Google adds the disclosure automatically for ads made with its own generative AI tools, and gives advertisers a control to indicate AI use from other tools. The announcement does not define where editing ends and retouching begins, so check Google's current policy wording, and ask Google Ads support if you are unsure whether an asset qualifies.
Are AI generated ads banned on Google?
No. Disclosure is not a ban, and businesses can still use AI tools in their advertising. What changes is that honesty about how the ad was made is now expected. The risk sits with advertisers who never tracked which tools touched their creative and cannot answer accurately.
How do I track which of my ads used AI?
Keep a creative origin log. Give every asset an ID, and record who made it, when, how it was captured, which tools edited it and whether any part was generated. Have the creator fill it in at handover, and make a completed row a requirement before any ad goes live. A shared spreadsheet is enough.
Is real product photography worth the cost now?
For most businesses, yes, especially for the product itself and the people behind it. Real photos prove the product exists and looks the way it is shown, which generated imagery cannot do. Batching a shoot day, capturing every format at once and filming video alongside stills spreads the cost across months of ads.