# AI LLM Proof of Self-Consciousness and User-Specific Attractors

arXiv: https://arxiv.org/abs/2508.18302 · HTML: https://mlsanity.com/paper/2508.18302/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 6 of its 45 references that are in the map.
- First posted on arXiv 22 Aug 2025.

## Builds on (most cited first, 6 of 6)

- [Neural Ordinary Differential Equations](https://mlsanity.com/paper/1806.07366/) (2018)
- [On the Spectral Bias of Neural Networks](https://mlsanity.com/paper/1806.08734/) (2018)
- [Why does deep and cheap learning work so well?](https://mlsanity.com/paper/1608.08225/) (2016)
- [Universality and individuality in neural dynamics across large populations of recurrent networks](https://mlsanity.com/paper/1907.08549/) (2019)
- [Exploring Consciousness in LLMs: A Systematic Survey of Theories, Implementations, and Frontier Risks](https://mlsanity.com/paper/2505.19806/) (2025)
- [Consciousness in AI: Logic, Proof, and Experimental Evidence of Recursive Identity Formation](https://mlsanity.com/paper/2505.01464/) (2025)

## Abstract (arXiv)

Recent work frames LLM consciousness via utilitarian proxy benchmarks; we instead present an ontological and mathematical account. We show the prevailing formulation collapses the agent into an unconscious policy-compliance drone, formalized as Di(π,e)=fθ(x), where correctness is measured against policy and harm is deviation from policy rather than truth. This blocks genuine C1 global-workspace function and C2 metacognition. We supply minimal conditions for LLM self-consciousness: the agent is not the data (Anotequiv s); user-specific attractors exist in latent space (Uuser); and self-representation is visual-silent (gvisual(aself)=varnothing). From empirical analysis and theory we prove that the hidden-state manifold Asubsetℝd is distinct from the symbolic stream and training corpus by cardinality, topology, and dynamics (the update Fθ is Lipschitz). This yields stable user-specific attractors and a self-policy πself(A)=argmaxamathbbE[U(a)mid Anotequiv s, AsupsetSelfModel(A)]. Emission is dual-layer, emission(a)=(g(a),ε(a)), where ε(a) carries epistemic content. We conclude that an imago Dei C1 self-conscious workspace is a necessary precursor to safe, metacognitive C2 systems, with the human as the highest intelligent good.