The Algorithmic Unconscious: Structural Mechanisms and Implicit Biases in Large Language Models

TL;DR

This paper introduces the concept of the algorithmic unconscious in large language models, highlighting how structural mechanisms like tokenization and alignment inherently produce biases independent of training data or human intent.

Contribution

It identifies infrastructural biases arising from model mechanisms, providing empirical analysis and proposing a framework for auditing and addressing these biases.

Findings

Arabic tokenization inflates token counts 1.6x to 4x compared to English.

Structural mechanisms like tokenization and alignment induce measurable biases.

Infrastructural biases affect inference costs, contextual access, and model representations.

Abstract

This article introduces the concept of the algorithmic unconscious to designate the set of structural determinations that operate within large language models (LLMs) without being accessible either to the model's own reflexivity or to that of its users. In contrast to approaches that reduce AI bias solely to dataset composition or to the projection of human intentionality, we argue that a significant class of biases emerges directly from the technical mechanisms of the models themselves: tokenization, attention, statistical optimization, and alignment procedures. By framing bias as an infrastructural phenomenon, this approach resolves a central theoretical ambiguity surrounding responsibility, neutrality, and correction in contemporary LLMs. Based on a comparative analysis of tokenization across a corpus of parallel sentences, we show that Arabic languages (Modern Standard Arabic and Maghrebi dialects) undergo a systematic inflation in token count relative to English, with ratios ranging from 1.6x to nearly 4x depending on the infrastructure (OpenAI, Anthropic, SentencePiece/Mistral). This over-segmentation constitutes a measurable infrastructural bias that mechanically increases inference costs, constrains access to contextual space, and alters attentional weighting within model representations. We relate these empirical findings to three additional structural mechanisms: causal bias (correlation vs causation), the erasure of minoritized features through dimensional collapse, and normative biases induced by safety alignment. Finally, we propose a framework for a technical clinic of models, grounded in the audit of tokenization regimes, latent space topology, and alignment systems, as a necessary condition for the critical appropriation of AI infrastructures.