(Inglês) Proceedings of 15th Meeting of Quality Researchers: Quality, Research and Development

Generative Artificial Intelligence (GenAI) has rapidly emerged as a transformative tool in scientific research. Large Language Models (LLMs) can produce human-like text and other content, offering new ways to generate ideas, review literature, formulate hypotheses, design studies, analyse data, and assist in scientific writing. These capabilities present significant opportunities to enhance creativity and efficiency across the research workflow. At the same time, the use of GenAI brings notable risks and challenges. Models often hallucinate incorrect information, reflect biases in training data, and raise concerns about academic integrity, privacy, and the potential deskilling or displacement of researchers. This paper provides a comprehensive yet concise overview of GenAI’s evolution and inner workings and examines its application in key stages of the scientific process. Each use-case is critically discussed, highlighting both the benefits and the pitfalls as reported in recent literature. A unified discussion of ethical and societal implications underscores the importance of responsible use, oversight, and policy in integrating GenAI into academia. The paper concludes by reflecting on prospects for GenAI in research and the need for guidelines that ensure these powerful tools are used to augment, not undermine, scientific integrity and progress.

--

Academic responsibility,Generative AI,Large language models,Research workflow

• Ciências da Computação e da Informação - Ciências Naturais