As seen in the December 2023 edition of The Advertising Club

By Shawn Spooner, CTO, billups

“All models are wrong, but some are useful,” British statistician George Box famously said back in the 1970s. In today’s world, where AI tools are automating tasks at a rapid pace, it is even more important to remember that while these technologies are helpful, they have their limitations.

Despite the rise of chatGPT and other generative AI tech, the fact is there are still serious limits to what it can do.

What does that mean for the world’s oldest form of advertising?

When planning an out-of-home ad strategy, data science and technology are powerful tools but must be managed with the creativity and strategic approach that comes only with the human touch. A strategy that blends human creativity with advanced machine learning is the best approach to creating engaging and effective OOH campaigns.

Can AI Do Anything? Considering the Limitations
AI is not limitless. It can only recombine what it already knows, and any OOH campaign must be created with that in mind.

All AI platforms are probability machines. If you can find the boundaries of their capabilities, you can break them. They are beholden to the data used to train them. Machine learning models are always different because of the randomization in how they are trained. Every model is unique, even with the same code and trained on the same data. Even when they behave the same way, it may be for different reasons, and it can be challenging to know why.

AI is great for telling you what works, but not at saying why it works. In certain ways, these machines are like airplane black boxes: They need a human to interpret the information. That said, there are certain things AI and advanced machine learning excel at — essentially, anything that can be pattern matched.

While the technology works fabulously in the case of automation — where a person tells the machine what to do, then it can do so on its own — the same tech struggles when there is a strategic element involved. What these systems are really doing is large-scale pattern matching. They can get arbitrarily complex and complicated in how they string facts together, but none of the models currently available are creative in the way humans are.

With the randomization of elements, AI can be used to generate creative, and eventually it will get good at that — but that is not true creativity. The model is simply repeating permutations until it comes up with something decent. But that is a combination, not creativity.

Beyond that, all advanced machine learning models have trouble encoding long-term sequences of relationships. As humans, we can understand long chains of causality, whereas AI models have finite time horizons. There is much they can learn, but they have limited capacity at present.

Another issue: It takes an enormous amount of computational power to make the tech work. Using such models is computationally expensive. To run a model as complicated as chatGPT, for example, takes a massive building and training process, as well as a lot of energy and expertise for building and maintaining the system. To build AI for OOH, it would need a purpose that makes sense in the industry.

AI Data Limitations in OOH
Another considerable limitation of using AI in OOH is that vast amounts of quality data must be available. AI is only as good as the data it is fed, but we don’t have such data at our fingertips because no one has been collecting and storing it digitally until more recently.

For someone to design a robust AI model for OOH, they would need clean data sources, and it is a burden to bring those together in a meaningful way.

While historical campaign data for the past couple of decades is not available, much of that data is now being collected. And while it’s too late for campaigns of yesterday to provide insight and analytics for ideal campaign scenarios, the future is moving toward that.

Essentially, as we collect data now for all campaigns, that information gets fed into the AI and will help us with more accurate planning in the future.

AI for OOH Creatives and Messaging

Can AI make creative?
While abundant AI software is available to create images, the quality of these images varies, and creativity is far limited to what was created before. Without human intervention, an AI graphic could easily be a copy or imitation of fast creatives that have been fed into the model. That could be an issue if an AI-generated content looks like a competitor’s past creatives, for example.

Though AI can generate ideas, relying upon it for your creatives in either art or content is not advisable.

AI for Campaign Planning 
Where AI can help, and help a great deal, is in campaign planning, where machine learning works with modeled data, analyzes that data, and learns from the results.

As the machinery is continuously evolving, it can suggest the best locations to meet goals and can increase the efficiency of the campaign plan.

Again, there are limitations. Obviously, we can ask ChatGPT about the goals and KPIs for an OOH campaign and get a very general answer. But if you get into specifics, the trustworthiness of the technology is called into question.

The bottom line: For AI to work in OOH campaigns, key variables like brand strategy and brand safety still very much require the human mind, and the many years of learning and expertise that come only with human experience.

‍