No Epistemic Respect for Bullshit Machines or LLMs

The excitement about Large Language Models (LLMs) has led some to consider the possibility that they could be artificial epistemic authorities. For instance, after considering arguments for and against granting LLMs the status of epistemic authorities, Hauswald (2025) argues that the standard view of epistemic authorities should be revised to accommodate LLMs as epistemic authorities. This paper aims to contribute to the literature on LLM or chatbot epistemology by sketching two arguments against granting LLMs the status of epistemic authorities. According to the first argument, if LLMs are bullshit machines, then they should not be granted the status of epistemic authorities that are worthy of our epistemic respect. According to the second argument, if LLMs cannot take responsibility and be held accountable for their outputs, then they should not be granted the status of epistemic authorities that are worthy of our epistemic respect.