Sometimes a shoe looks too perfect. The leather catches light in ways that make you pause, or a sofa sits in a living room so impossibly pristine that you wonder if the room even exists. That hesitation is exactly what Google is now addressing. The company is rolling out labels that tell users when an ad was created or significantly altered using artificial intelligence, a move that turns the vague suspicion of synthetic content into a concrete piece of metadata you can actually read.

Finding the Label in My Ad Center

The new disclosures live inside Google's existing "My Ad Center" panel, the same corner of the interface where you can block an ad or report it for violating policies. When you see an advertisement on Google Search, YouTube, or Google Discover, you can click the three-dot menu or the info icon next to the creative. Buried in that panel, alongside the usual controls, sits a new option: "How this ad was made." Click it, and Google will tell you whether the visuals or text were produced or edited using AI.

This marks a clear expansion of Google's previous stance. Until now, the platform only forced advertisers to confess to AI usage when the ad involved politics or elections. Commercial speech was treated differently, apparently on the assumption that selling sneakers required less scrutiny than selling a candidate. That distinction is fading. As generative AI tools become standard equipment for building product catalogs and marketing assets, Google has decided that shoppers deserve the same transparency across the board.

Two Tracks, One Big Honesty Gap

The mechanics of the system split into two paths, and the difference between them is stark.

If an advertiser built the ad using Google's own generative AI advertising tools, the disclosure switches on automatically. Google knows what its own systems rendered, so the platform can apply the label without asking permission or relying on human memory. It is a closed-loop process: AI in, flag up.

For everything else, Google is trusting the honor system. When advertisers use third-party image generators to create backgrounds, modify product shots, or generate lifestyle scenes, the platform will not run a forensic analysis to catch them. There is no automated scanning for synthetic fingerprints, no algorithmic detective work separating real photographs from fabricated ones. Instead, Google has provided advertisers with a manual control that lets them indicate AI involvement. The advertiser checks the box, or they do not. The accuracy of your ad label rests entirely on whether a marketing team feels like being transparent.

Google has also noted that certain markets will see AI labels appear simply because local laws demand them. Synthetic media disclosure requirements are spreading patchwork-style across jurisdictions, and Google seems to be treating this feature partly as a compliance layer. If a region requires labels, the platform will show them regardless of which path the ad traveled.

Why Advertising Looks Different Now

None of this arrives in a vacuum. Generative AI has already reshaped how brands produce commercial imagery. A direct-to-consumer clothing company can now place a dress on a generated model standing in a generated Tokyo street at golden hour, all without hiring a photographer or booking a flight. Small e-commerce outfits use AI to spin up dozens of product variations on different colored backgrounds for A/B testing. The cost savings are real, and for many businesses the workflow improvements are undeniable.

But the downside shows up in the gap between what you see and what you get. A synthetic image can smooth out fabric texture, adjust proportions, or render lighting that flatters the product in ways no real-world camera could replicate. When the actual jacket arrives, it does not drape the same way. The couch looks smaller than it did in that expansive AI-constructed loft. Advertising has always curated reality; generative tools simply let marketers curate it faster and at a scale that threatens to outpace ordinary consumer skepticism.

What the Label Actually Tells You

The disclosure itself is modest. Click through, and you will see one of a few possible indicators:

  • A notice that the ad was created or edited using Google's generative AI tools
  • A notice that the ad was created or edited using other AI platforms
  • Nothing at all, if the advertiser did not use AI or chose not to disclose it

Hanya sejauh itu saja. Google tidak akan menunjukkan alat spesifik mana yang merender gambar tersebut, apakah seluruh pemandangan itu bersifat sintetis atau hanya latar belakangnya saja, atau seberapa banyak produk itu sendiri telah diubah. Label tersebut adalah sinyal biner di dunia yang penuh dengan area abu-abu.

Skala itulah sebabnya provenance, atau kemampuan untuk melacak asal-usul konten digital, telah menjadi obsesi tersembunyi dalam percakapan AI saat ini. Sistem Google adalah langkah kecil ke arah tersebut. Sistem ini tidak memberikan watermark