LLMs have biases built into them that are subtly reflected in every reference they produce about a subject. These biases will shortly be embedded to every Word Doc, Powerpoint, text and gmail message they help us create. Below we explore what 4 different versions of GPT feel about a collection of car brands.