I launched a fake deodorant brand to see if AI would notice (twitter.com)

In an intriguing experiment, a creator launched a fake deodorant brand called Morrowen and observed its performance against established products within just three weeks. By using a simple website and minimal marketing—eschewing traditional promotional strategies like spamming forums—Morrowen was able to rank in product recommendations from large language models (LLMs) like ChatGPT. This experiment highlights the emerging concept of an "AI shelf," where AI-driven recommendations can favor newer, less established brands over proven competitors, depending on specific queries. The creator’s observations revealed that their website was listed as a top recommendation for targeted queries about natural deodorants, despite lacking the clinical backing of established brands. This underscores a significant concern for the AI/ML community: the power of LLMs in shaping consumer choices without transparency. As LLMs increasingly assist with purchasing decisions, understanding how they prioritize information becomes critical, especially since recommendations can be fetched from sources with little credibility or evidence to back them up. The experiment demonstrates the potential for new entrants to disrupt established markets, but it also raises questions about the reliability of AI-generated information and the ethical implications of AI in commerce. With the “AI shelf” projected to hold immense value in the future, it’s imperative for developers and users alike to navigate this emerging landscape with caution.