{"ID":2880314,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2508.14802","arxiv_id":"2508.14802","title":"Privileged Self-Access Matters for Introspection in AI","abstract":"Whether AI models can introspect is an increasingly important practical question. But there is no consensus on how introspection is to be defined. Beginning from a recently proposed ''lightweight'' definition, we argue instead for a thicker one. According to our proposal, introspection in AI is any process which yields information about internal states through a process more reliable than one with equal or lower computational cost available to a third party. Using experiments where LLMs reason about their internal temperature parameters, we show they can appear to have lightweight introspection while failing to meaningfully introspect per our proposed definition.","short_abstract":"Whether AI models can introspect is an increasingly important practical question. But there is no consensus on how introspection is to be defined. Beginning from a recently proposed ''lightweight'' definition, we argue instead for a thicker one. According to our proposal, introspection in AI is any process which yields...","url_abs":"https://arxiv.org/abs/2508.14802","url_pdf":"https://arxiv.org/pdf/2508.14802v1","authors":"[\"Siyuan Song\",\"Harvey Lederman\",\"Jennifer Hu\",\"Kyle Mahowald\"]","published":"2025-08-20T15:52:34Z","proceeding":"cs.AI","tasks":"[\"cs.AI\",\"cs.CL\"]","methods":"[\"Large Language Model\"]","has_code":false}