Skip to content

AI_USAGE.md

Latest commit

 

History

History
37 lines (26 loc) · 2.84 KB

AI_USAGE.md

File metadata and controls

37 lines (26 loc) · 2.84 KB
Raw

AI Usage Declaration

This file declares the usage of Artificial Intelligence (AI) and Large Language Models (LLMs) applications in the development of this project.

1. Tools Applied

  • [Tool Name, e.g., GitHub Copilot] – Applied primarily for [e.g., inline code completion].
  • [Tool Name, e.g., OpenAI GPT-4] – Applied primarily for [e.g., code alteration, text restructuring, error localization].

2. Generative Usage

This section catalogs instances where AI was used to generate text, code, or data directly from prompts, without a pre-existing human draft.

  • Code Boilerplate: [e.g., "Generated the initial class structures and skeleton code for the data ingestion pipeline in src/ingest.py."]
  • Unit Tests: [e.g., "Generated foundational unit tests for the utility functions in tests/test_utils.py."]
  • Documentation Drafting: [e.g., "Generated the initial structure of the Sphinx docs/ folder."]

3. Alterative Usage

This section catalogs instances where human-authored content or code was submitted to an AI system to be restructured, reformatted, or translated.

  • Code Alteration: [e.g., "Applied AI to restructure the nested loops in the core mathematical model (src/model.py) into vectorized operations."]
  • Error Localization: [e.g., "Input stack traces into the AI to locate the source of memory leaks during HPC cluster deployment."]
  • Text Alteration: [e.g., "Applied AI to alter the phrasing, syntax, and grammar of the methodology documentation and README."]
  • Data Reformatting: [e.g., "Applied AI to generate scripts that convert raw instrument outputs from .txt format into structured .json format."]

4. Strict Exclusions

The application of AI tools was strictly excluded from the following areas of the project:

  • Hypothesis & Experimental Design: The core scientific questions, theoretical frameworks, and experimental methodologies were developed entirely by the human researchers.
  • Data Manipulation: AI tools were not used to synthesize, alter, interpolate, or extrapolate raw experimental or observational data.
  • Result Interpretation: The analysis of pipeline outputs and the formulation of scientific conclusions were performed solely by human authors.

5. Human Oversight & Verification

This section details the protocols applied to review AI-generated and AI-altered outputs prior to integration.

  • Testing: All AI-generated and AI-altered code underwent human review and was required to pass the automated test suite (pytest) prior to merging.
  • Literature Verification: Algorithmic structures or scientific concepts output by the AI were cross-referenced with primary literature.
  • Responsibility: The human contributors hold absolute and final responsibility for the accuracy, legality, and scientific validity of all code, data, and text within this repository.