Google Enlisted Three Advertising Legends to Make AI Creative Respectable

Google unveiled the finished work from The Small Brief on August 19. Three of advertising’s most decorated creative directors, given unlimited access to Google’s AI video tool, spent it on a Hudson River Valley farm, a Long Island ferry and a nonprofit for caregivers.

Susan Credle made a campaign for Stonewood Farm about regenerative agriculture. Tiffany Rolfe built a piece for South Ferry that visualizes two centuries of a service founded in 1833. Jayanta Jenkins made a film for ARCHANGELS, which supports unpaid caregivers.

The work is for small businesses. The campaign is not.

Ad Age read the project the same way, asking in its headline whether the spots will help ease the stigma around AI creative, and describing the strategy elsewhere as Google’s move to make AI likable by enlisting star creatives. That’s the transaction. Google supplied compute. Three people supplied 30 years of reputation each.

What the creatives actually said

None of them oversold it. Rolfe, global chair and global chief creative officer at R/GA, described the state of the industry as unresolved:

“We’re all adopting it, but at the same time we’re all a bit hesitant. The more we start to push it, not to replace the things we used to do but to unlock new things, we’ll understand what it can really do for us.”

That’s a hedge, and the honest one. “We’re all a bit hesitant” is not the sentence a vendor writes for a testimonial. It made the cut anyway. Google understood that a polished endorsement from these three would’ve read as bought and landed as nothing.

Why a company that owns the tool needs other people to vouch for it

The reason is a trust problem Google can’t solve by shipping features.

We reported last week on the size of it: ad executives believe younger audiences feel good about AI-made advertising far more often than those audiences actually do, and the gap has been widening, not closing. Brands have started treating that gap as a positioning opportunity rather than a problem to manage, which is what Liquid Death did by making an ad against AI three days ago.

When “not made with AI” becomes a selling point, a company selling AI creative tools has a category problem, not a product problem. Features don’t fix category problems. Credibility does, and it has to be borrowed from people who already hold it.

Hence the structure of this campaign. Not a demo reel. Three named humans with reputations to lose, working on charity-adjacent briefs where the motive reads as generous, published with their doubts intact.

The number that undercuts the whole premise

Buried in Ad Age’s coverage is research from OpenAI on how marketers use these tools. Creative ideation accounts for 1.4% of marketers’ ChatGPT usage.

1.4%. Not a rounding error on the way to something bigger, a ceiling on where the tools are being trusted.

The rest is the unglamorous middle of the job. And the same research found the traffic runs both ways across functions: engineers put about 20% of their cross-function AI use into marketing tasks, while marketers spend about 17% of theirs on engineering work.

So the campaign argues that AI unlocks new creative possibility, while the usage data says the tools are mostly being used to do admin faster. Both can be true. But a marketing leader watching this campaign should notice that the case for AI creative is being made with borrowed reputations rather than with adoption numbers, and ask why the adoption numbers weren’t the pitch.

The steelman

The strongest argument for Google here: the work exists, it went to organizations that could never have afforded film production at this level, and a 200-year history is hard to shoot and easy to generate. South Ferry cannot send a camera crew to 1833.

That’s real, and it’s the best case for generative video anyone has made this year. The work displaced nothing. It reached clients who’d never have been served at all.

It also doesn’t answer the trust question. Audiences who dislike AI advertising won’t like it more because the brief was charitable. If anything, a nonprofit spot that turns out to be generated carries more downside than a soft drink one, because the expectation of sincerity is higher.

What to take from it

If you’re deciding how to talk about your own AI use, this campaign is both a template and a warning.

The template: name the people, show the doubt, pick work that would not have existed otherwise. Google let its participants say they were hesitant, which is why the campaign reads as a conversation rather than an ad.

The warning: Google needed three borrowed reputations to make this argument, and it’s the biggest advertising company on earth. If that’s the cost of making AI creative respectable at Google’s scale, a mid-sized brand announcing its AI creative workflow in a LinkedIn post has badly underpriced the problem.

The safest position is still the one the usage data supports. Use it for the 98.6% nobody writes campaigns about, and stay quiet about the 1.4% until your audience tells you they’re comfortable.