Go to a language model, Claude, Gemini, ChatGPT, or any other, start a new conversation and ask this question:

Give me a number between 1 and 10.

What did you get? A seven?

The first time I heard about this experiment was during the keynote of Zoe Scaman, strategist and founder of Bodacious. And this example stuck with me because it illustrates what large language models like ChatGPT, Gemini, and Claude are: they are prediction machines. They are not giving you the best answer, the truth, a fact, or whatever. All they give you is the most probable answer.

Now, as many of us increasingly rely on the same data that is easily accessible through LLMs, our output is starting to look the same. 

A large meta-study, titled “Does Generative AI Make Us Think Alike?: A Systematic Review and Meta-Analysis of Homogenization Effects in Human–AI Co-creation”, was recently published by Tilburg University It covered 19 studies and 61 effect sizes. One of their most striking findings was that AI-assisted outputs (depending on the task) tend to become more similar to each other than outputs produced without AI.

The article described how the algorithms behind the current models might actually lead to more of the same in the future. They called this phenomenon the sociotechnical feedback loop, and it goes like this: AI learns from human behaviour, humans then adapt to AI’s suggestions, and over time both start reinforcing the same ideas and patterns. This can make outputs increasingly similar.

If we increasingly rely on the same tools for output, how can we still differentiate and appear authentic as a brand?

This brings us back to Zoe and other research like that of Gartner, which shows that the most valuable assets in terms of differentiation today are your talent, workflows and culture, or as Zoe describes it: your heartwood.

The heartwood of a tree is the non-visible, slow-growing, and dense inner core that provides strength and structure to a tree. The metaphor was used by Zoe to describe the organic and deeply rooted human knowledge within an organization or community that cannot be replicated by artificial intelligence.

As AI strategist Craig Hepburn puts it:

“A model only knows what it was trained on, and it was never trained on you. The way your business actually works, the data sitting in your systems, the customers you have known for years, the hard facts of your particular corner of the world, were never in the data it learned from and never will be.”

It’s that type of data that enables brands to distinguish themselves, and it might be worth exploring, defining, understanding, and cultivating further.