The advertising industry is under pressure to clean up its act and there are legitimate concerns about Generative AI and its carbon footprint. Nick Hunter, the CEO and co-founder of Paper Moose, argues that such fears are misplaced, and in the creative process, AI technology can help reduce carbon emissions, rather than accelerate it.
Glance at the headlines surrounding AI at the moment, and on top of job-replacement fears, there’s an equal concern LLMs are about as environmentally friendly as a whale-powered paper mill. The concern is valid, and as an industry, we need to take responsibility for every watt we use. However, such concerns also gloss over context and intent, while missing AI’s hidden potential as a form of “creative darning”, or upcycling, yielding greater longevity from your creative carbon.
Training large AI models is energy-intensive. This cannot be ignored. In practical agency use, however, most generative creative work is done with trained models, where the act of generating and refining assets consumes a fraction of the energy required by traditional methods.
To go data-led for a moment (I know how much we all love stats), the average 30-second TV commercial, shot on location with a full crew, typically produces anywhere from 1 to 2 tonnes of CO₂ in production emissions alone. That figure accounts for lighting, power on set, crew and talent travel, prop and set construction, catering and waste. Scale that up across multiple markets or campaign refreshes, and the resource drain grows rapidly.
By comparison, generating an equivalent 30-second video sequence with state-of-the-art AI models in a renewable-powered data centre typically uses around 15 to 40 kilowatt-hours. That equates to just 5 to 18 kilograms of CO₂, even when factoring in current global energy mixes, that’s as little as 1 or 2 percent of the carbon output of a conventional shoot.
It is also worth noting that the environmental impact of actually airing that 30-second ad can far exceed either method of production. Media buying and distribution, especially across digital platforms and broadcast television, account for the greatest energy consumption in the content value chain. In this context, the responsibility to drive efficiency at every stage, from concept to screen, cannot be overstated.
But with generative AI, we finally have the means to adapt, extend, and refresh assets again and again, making campaigns more resilient and sharply reducing waste.
Recently, in our own studio, we put this into practice. One client’s product line changed just prior to launch. In the past, we would have lost the campaign and reshot the ad from scratch. Instead, generative AI allowed us to modify the 30-second spot digitally, updating products in every frame with zero need to ship the products to location, fly in talent, or reshoot anything.
The refreshed work was indistinguishable from a full reshoot, except its carbon footprint had dropped by over 95 percent. In this way AI can become the great recycler; instead of trashing campaigns once they have had their moment in the sun, we can find new ways to refresh with minimal environmental and bottom line cost.
Of course, I am not proposing that live action production disappears. There will always be a place for beautifully crafted original footage, high-touch storytelling, and on-location work.
Used with clear intent and creative discipline, AI is not a device for endless low-value content. When every decision from storyboard to impression has an environmental cost, smart industry leaders must leverage every tool to keep that cost as low as possible, without sacrificing creativity. In this way, AI can and should be seen as a tool that gives existing creative ideas an easier pathway to constantly contemporary relevance, and thus sustainability and longevity.
Nick Hunter is the CEO and co-founder of Paper Moose.
