Abstract

Artificial intelligence tools based on Large Language Models (LLMs), such as OpenAI’s ChatGPT, are revolutionizing how we process and synthesize information. However, these tools can be anemic when applied to niche or unconventional domains, such as parapsychology. This is due to the fact that the input content used to train these LLMs draws predominately from mainstream internet sources (e.g., Wikipedia), and thus, the corresponding outputs inherit various biases surrounding heterodox subjects. This review introduces UnityGPT, a custom AI tool provided with a vast array of unconventional research from a plurality of disciplines, intended for use as a primary research tool in the frontier sciences. The article summarizes the technology approach of both ChatGPT and UnityGPT, and their relative strengths and weaknesses.