DOI of the published preprint https://doi.org/10.1590/1678-98732433e018
Towards Ethical, Responsible, and Human-Centered Academic Writing with AI
DOI:
https://doi.org/10.1590/1678-98732433e018Keywords:
academic writing, scientific integrity, Generative Artificial Intelligence, authorship, narrative reviewAbstract
Introduction: The rapid spread of generative artificial intelligence (GenAI) has transformed academic writing, both enhancing efficiency and prompting concerns about authorship, integrity, privacy, and cognitive development. Although recent scholarship has focused largely on normative debates and institutional policies concerning authorship and transparency, it offers little practical guidance on how to use AI without compromising intellectual formation, particularly for early-career researchers. This article addresses that gap by proposing a structured method for the ethical, responsible, and human-centered use of AI. Materials and methods: The study conducts a narrative review of the literature examining the risks associated with AI-assisted writing, such as diluted authorship, cognitive offloading, bias, privacy concerns, and “hallucinations”, and integrates this review with normative documents and institutional guidelines. Building on this synthesis, the article develops a practical guide organized into five methodological phases designed to foster critical and responsible engagement with AI technologies. Results: The proposed guide structures the writing process into five stages: initial human-centered planning; AI-assisted drafting paired with critical curation; authorial appropriation and rewriting; rigorous verification of facts, sources, and argumentative coherence; and final accountability through transparent disclosure of AI use. Each stage provides safeguards to mitigate cognitive, ethical, and epistemic risks while preserving authorial agency. Discussion: The argument advanced here is that ethical AI use depends less on formal prohibitions and more on cultivating practices of intellectual vigilance, meticulous documentation, and continuous reflection. While the proposed process may limit some of the efficiency gains associated with AI, it strengthens cognitive capacities, safeguards scientific integrity, and helps balance technological innovation with human responsibility.
Downloads
Posted
How to Cite
Section
Copyright (c) 2025 Rafael Sampaio

This work is licensed under a Creative Commons Attribution 4.0 International License.
Plaudit
Data statement
-
The research data is contained in the manuscript


