Nearly every advertising category is currently having some version of the same anxious conversation about generative AI: how much of this work can be automated, how quickly, and what happens to the people currently doing it. Most creative formats have already received at least a partial answer, and for the humans involved it has rarely been reassuring. Product photography, voiceover, script writing, and increasingly the on-camera talent itself have all proven susceptible to synthetic replacement at a fraction of the cost.
Street Poller Media founder Shane Ginsberg believes the unpredictability and authenticity of street interviews will be the last frontier for generative AI to crack.
His argument starts with a direct comparison. A user-generated content creator filming a product review from their kitchen is, in Ginsberg’s assessment, a scene AI can already fake convincingly. The lighting is controlled, the background is static and generic, the performance is scripted, and even the person delivering the pitch can now be synthesized well enough that most viewers would not detect the substitution. Every variable in that shot is one an AI model can reproduce because every variable was already artificial.
A street interview presents a fundamentally different problem. The specific streetlights illuminating the shot, the actual license plates on cars passing behind the interviewer, the particular quality of daylight at that hour in that city, and the genuine spontaneity of what happens when a second stranger wanders unexpectedly into frame or reacts to something unprompted, none of that is currently replicable. Ginsberg has argued it may be among the last things AI becomes capable of faking convincingly, if it ever fully manages it. Simply put, the value of the footage depends on it being an unrepeatable moment between real people, and a synthesized version is by definition not that.
The distinction is not merely philosophical for the business. It forms the basis of an economic argument about where creative value accrues as synthetic content becomes cheaper and more widespread.
The logic runs roughly as follows. As AI-generated marketing content becomes trivially cheap to produce, formats that AI can replicate should get commoditized, because supply expands toward infinity and the marginal cost of another asset approaches zero. Anything competing purely on production volume loses its scarcity. Formats that AI cannot replicate should, by the same logic, hold or increase in value, because they retain scarcity in an environment where nearly everything else has lost it.
Street interview advertising, in Ginsberg’s framing, sits firmly in the second category. Its entire value proposition depends on something no current model can synthesize: verifiable evidence that a real interaction occurred between a real interviewer and a real stranger in a real public place.
That thesis intersects with something happening on the consumer side as well. Ginsberg has argued that audiences are growing fatigued with AI-generated marketing content, and notably, not primarily because it is poorly made. The fatigue, in his view, stems from the fact that synthetic content does not connect the way an authentic unscripted conversation does, and no amount of additional production polish resolves that, because the problem is not quality. The problem is whether a viewer believes what they are watching actually happened.
If that read is correct, the competitive dynamics of the next several years may look different than the current automation narrative suggests. Rather than AI displacing human-made advertising wholesale, the market may bifurcate, with synthetic content dominating anything where authenticity is not the point, and formats that can prove human reality commanding a premium precisely because they can no longer be faked.
Ginsberg has taken that reasoning to what he sees as its logical conclusion, arguing that street polling will end up creating a substantial number of jobs over the coming years, at a moment when much of the surrounding economy is moving in the opposite direction.
The post The One Creative Format AI Can’t Replace Yet With Shane Ginsberg of Street Poller Media first appeared on Tycoonstory Media.
Source: Cosmo Politian
