Do Large Language Models Hallucinate Electric Fata Morganas?

A new arXiv paper (arXiv:2608.18816v1) argues that AI hallucinations in large language models have philosophical significance for machine consciousness, presenting two empirical investigations. The authors found that higher temperature settings in GPT models produce plausible but incorrect answers, while lower temperatures yield factually accurate ones, and that hallucinations stem from exposure to subjective training data rather than cognitive ability. The paper claims that a model's self-reports of emotion or sentience fall within the definition of hallucination, making machine consciousness potentially epistemically inaccessible.

arXiv:2608.18816v1 Announce Type: new Abstract: AI hallucinations - that is, outputs which are made up, cannot be verified, or contradict the source material - are generally regarded as an engineering flaw to be dealt with. This paper contends that they also have philosophical significance when it comes to the question of machine consciousness. We examine the known causes of hallucinations in large language models - such as source-target divergence, discrepancies between training and inference, and overfitting - and we present two empirical investigations. In the first, we apply successive generations of the GPT model to ambiguous factual questions under different temperature settings, finding that higher temperatures result in plausible but incorrect answers while lower temperatures lead to factually accurate ones. The sampling parameters that cause a model to seem creative or spontaneous and thus more likely to pass behavioral tests of intelligence are the same ones that increase its hallucination rate. In the second, we look at an encoder-only model that has been trained on encyclopedic data and which answers questions of the same type factually and without embellishment, indicating that hallucinations are due to exposure to subjective and socially diverse training data rather than to the development of any cognitive ability. Using references to Turing, Searle's Chinese Room, the frame problem, and the cybernetic tradition of Wiener and Ashby, we claim that a model's self-reports of emotion or sentience come within the definition of hallucination, and that any future occurrence of machine consciousness might remain epistemically inaccessible since it would be indistinguishable from a sufficiently advanced hallucination.