Google has started rolling out a How this ad was made label that shows when an ad was created or edited with AI. It does not ban AI creative, it just puts a tag on it, which means lazy machine output now stands out for the wrong reasons.
Disclosure does not kill AI creative. It just moves the question from whether you used it to whether it was any good.
Google will now tell people when an ad was made with AI. A new option called "How this ad was made" shows whether an ad was created or edited using AI tools. It started rolling out in the second week of July.
The label sits alongside the ad and gives users a way to see the machine's fingerprints. It lands as generative tools move from novelty to default inside ad production, with a growing share of creative now touched by AI at some stage. Google is not banning any of it. It is putting a tag on it.
Why it matters
Australian marketers leaning on generative tools for speed now have to assume the work can be labelled as such. That is not a reason to stop. It is a reason to make sure the output earns its place. A cheap AI ad with a disclosure tag on it looks worse than a cheap ad ever did on its own.
Trust is the real currency here. Consumers are already sceptical of what they see online, and regulators here have been circling misleading advertising and AI-generated content for a while. A label that says "made with AI" will make some people look twice, so the creative has to hold up under that second look.
There is an upside for anyone doing the work properly. If the market gets flooded with obvious machine output, human craft becomes a point of difference again. The tag punishes lazy production and rewards the brands that still care how the ad reads.
Google pulled in around US$264 billion in advertising in 2024, and it will now flag which of those ads were built with AI
What to do about it
Assume any AI-assisted ad can be labelled and judge it on whether it still persuades.
Put human editing on top of AI drafts so the work does not read as machine filler.
Keep your brand voice tight. Disclosure makes generic AI output stand out for the wrong reasons.
Check your claims. AI drafts invent things, and a labelled ad with a dodgy claim is a compliance risk here.
Watch performance on labelled versus unlabelled creative once the tag is widespread. Let the data tell you if it moves behaviour.
The tools are not the problem. Lazy work with a label on it is.