The Large Language Models and Ethical Issues in Scientific Research

Abstract

In the current information age, rapid advances in Artificial Intelligence (AI) and Machine Learning (ML) are transforming countless disciplines. This dissertation explores the emerging role of Large Language Models (LLMs), such as GPT-4 and AutoGPT, as innovative tools that can generate human-like text and enhance scientific research processes. Positioned at the intersection of computational efficiency and intellectual enquiry, LLMs offer unparalleled opportunities to automate labour-intensive tasks traditionally associated with scientific research. While these models have the potential to reshape conventional research paradigms and adapt rapidly to the evolving demands of the discipline, their use introduces complex ethical dilemmas. These range from algorithmic bias and misinformation to concerns about data privacy and transparency, and the erosion of accountability in scientific results. This dissertation aims to critically evaluate the capabilities and limitations of LLMs in scientific research and explore their potential for responsible integration into research workflows. The study includes a scoping review, conducted following PRISMA guidelines. Multiple online databases, such as SCOPUS, ERIC, EBSCO, and Web Of Science (WoS), were searched for relevant literature, focusing on studies that utilize Large Language Models in scientific contexts. A survey was conducted among researchers from INESC-TEC and Master's students involved in research activities utilizing LLMs, to identify opinions of the current applications, ethical challenges, and best practices regarding LLMs in research. The survey collected qualitative and quantitative data on research aims, findings, and ethical considerations related to LLM use. While Large Language Models hold significant potential to advance scientific research, their responsible integration necessitates careful ethical considerations. This study highlights the positioning of researchers in earlier and more advanced areas of scientific research in relation to the use of LLMs in their research activities.