Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Architectures of Error: A Philosophical Inquiry into AI and Human Code Generation

Camilo Chacón Sartori

TL;DR

This paper addresses the epistemic distinction between GenAI-generated and human-generated code by introducing Architectures of Error, a four-dimensional framework spanning semantic coherence, security robustness, epistemic limits, and control/debugging. It grounds the analysis in Dennett's mechanistic functionalism, Rescher's methodological pragmatism, and Floridi's Levels of Abstraction to explain why AI and human code generation exhibit fundamentally different error origins and implications for verification and trust. The main contributions are a principled taxonomy of AI- and human-originated errors, guidance for architecture-aware verification and validation, and a pragmatic approach to responsibility and trust calibration in human–AI software development. The findings have practical significance for designing robust V&V processes, governance structures, and collaboration practices that recognize the distinct limitations and strengths of GenAI tools in real-world software engineering.

Abstract

With the rise of generative AI (GenAI), Large Language Models are increasingly employed for code generation, becoming active co-authors alongside human programmers. Focusing specifically on this application domain, this paper articulates distinct ``Architectures of Error'' to ground an epistemic distinction between human and machine code generation. Examined through their shared vulnerability to error, this distinction reveals fundamentally different causal origins: human-cognitive versus artificial-stochastic. To develop this framework and substantiate the distinction, the analysis draws critically upon Dennett's mechanistic functionalism and Rescher's methodological pragmatism. I argue that a systematic differentiation of these error profiles raises critical philosophical questions concerning semantic coherence, security robustness, epistemic limits, and control mechanisms in human-AI collaborative software development. The paper also utilizes Floridi's levels of abstraction to provide a nuanced understanding of how these error dimensions interact and may evolve with technological advancements. This analysis aims to offer philosophers a structured framework for understanding GenAI's unique epistemological challenges, shaped by these architectural foundations, while also providing software engineers a basis for more critically informed engagement.

Architectures of Error: A Philosophical Inquiry into AI and Human Code Generation

TL;DR

This paper addresses the epistemic distinction between GenAI-generated and human-generated code by introducing Architectures of Error, a four-dimensional framework spanning semantic coherence, security robustness, epistemic limits, and control/debugging. It grounds the analysis in Dennett's mechanistic functionalism, Rescher's methodological pragmatism, and Floridi's Levels of Abstraction to explain why AI and human code generation exhibit fundamentally different error origins and implications for verification and trust. The main contributions are a principled taxonomy of AI- and human-originated errors, guidance for architecture-aware verification and validation, and a pragmatic approach to responsibility and trust calibration in human–AI software development. The findings have practical significance for designing robust V&V processes, governance structures, and collaboration practices that recognize the distinct limitations and strengths of GenAI tools in real-world software engineering.

Abstract

With the rise of generative AI (GenAI), Large Language Models are increasingly employed for code generation, becoming active co-authors alongside human programmers. Focusing specifically on this application domain, this paper articulates distinct ``Architectures of Error'' to ground an epistemic distinction between human and machine code generation. Examined through their shared vulnerability to error, this distinction reveals fundamentally different causal origins: human-cognitive versus artificial-stochastic. To develop this framework and substantiate the distinction, the analysis draws critically upon Dennett's mechanistic functionalism and Rescher's methodological pragmatism. I argue that a systematic differentiation of these error profiles raises critical philosophical questions concerning semantic coherence, security robustness, epistemic limits, and control mechanisms in human-AI collaborative software development. The paper also utilizes Floridi's levels of abstraction to provide a nuanced understanding of how these error dimensions interact and may evolve with technological advancements. This analysis aims to offer philosophers a structured framework for understanding GenAI's unique epistemological challenges, shaped by these architectural foundations, while also providing software engineers a basis for more critically informed engagement.

Paper Structure

This paper contains 23 sections, 1 figure.

Table of Contents

  1. Introduction
  2. Comparative Analysis of Functional Competences and Failure Modes
  3. Dimension 1: Semantic Coherence and Verification Challenges
  4. Semantic incoherence in GenAI
  5. Dimension 2: Security Robustness and Risk Profiles
  6. Dimension 3: Epistemic limits
  7. Context Integration
  8. Adaptability and Generalization Boundaries
  9. Dimension 4: Control Mechanisms and Debugging
  10. Output Consistency
  11. Process Traceability
  12. Dimensions interactions
  13. Functional Consequences for Software Engineering Practice
  14. Philosophical Challenges in Verifying and Validating GenAI-Generated Code
  15. The Attribution Problem: Rethinking "Responsibility"
  16. ...and 8 more sections

Figures (1)

  • Figure 1: The causal progression from the foundational architecture (GenAI and Cognitive) to specific error modes (dimensions) and their overall implications.