Still Human: Why Consumer Closeness Matters in an AI-Driven World.
There has been a lot of buzz around the generation of Synthetic Data and the direct extraction of Insight from Large Language Models (LLMs). And while there is a lot of potential in these approaches, there are also serious pitfalls.
The nature of an LLM is to produce plausible output. Another way of saying this is that LLMs are fantastic at producing stereotypes based on their training.
LLMs are powerful tools, and they can be used to help us generate Insight. But the true power of these models is not simply asking them to give us (plausible) answers, but rather to bring real data, from real humans, to them and use their power to help us better understand it.
You can read here a thought-piece on this topic I wrote for MarTech Magazine called Still Human: Why Consumer Closeness Matters in an AI-Driven World.