AI-generated content is becoming increasingly common in marketing, offering fast and low-cost solutions. However, a recent analysis indicates this may come at the expense of creative quality. A study conducted by global advertising firm TBWA and Monash University found that AI systems often fall short when attempting to replicate highly innovative campaigns originally developed by humans.
The research team chose 1000 standout advertising concepts and processed them through large language models (LLMs), providing only a one-sentence summary for each. Expectations were high that AI might match or even enhance the creative brilliance found in human efforts. Instead, the LLMs consistently removed the unique qualities of the campaigns, such as metaphor, humour and imagination. These elements have long been considered central to creative success.
Analysts at TBWA observed a clear trend. While the AI-generated content appeared polished and informative, it lacked originality. Even when different versions of LLMs were used, results did not improve. This suggests that data processing and computational power alone do not translate into imaginative thinking. For an industry aiming to use AI to increase productivity, this is a notable challenge.
The findings come as many major brands explore AI-generated advertising to cut costs and increase efficiency. Yet, if AI continues to prefer predictability over creativity, marketers may end up producing campaigns that fail to stand out. Experts note that these tools operate on patterns of probability, not innovation, which can make it harder to create a unique brand voice.
TBWA plans to continue research into how consumers respond to content they know is created by AI. While automation can help streamline processes, it appears that human input is still essential when it comes to achieving creativity and distinctiveness.
