Chatting with Pythons: using ChatGPT and Python computer code to screen abstracts for systematic literature reviews in complex disciplines

Author:

Mudd Alexandra1,Conroy Tiffany2,Voldbjerg Siri3,Goldschmied Anita1,Schuwirth Lambert2

Affiliation:

1. Sheffield Hallam University

2. Flinders University

3. Aalborg Universitet

Abstract

Abstract

Literature reviews are essential to scientific research; but abstract screening is time intensive, subject to human error and is a non-creative task, arguably impeding research development. Publicly available generative AI is revolutionising human activity creating new opportunities alongside new challenges for academic research. Studies have illustrated the power of AI in abstract screening for discrete exact science projects. However, individual researchers or small research teams, may lack the skills to independently use AI tools in a systematic, reproducible, transparent, and expandable manner. Also, AI’s capabilities to explore complex human activity is unclear. Here we demonstrate ChatGPT’s abilities to screen abstracts examining complex problems intersecting education, language, and human activity, in a manner that is systematic, reproducible, and transparent. We further demonstrate how coding-naïve researchers can use publicly available education in AI prompt engineering and Python computer coding using ChatGPT as a knowledgeable companion alongside assistance from interdisciplinary colleagues to use ChatGPT effectively. The result is an efficient first-line screening tool for academic researchers. We share our publicly available Python computer coding script with interdisciplinary colleagues to facilitate their exploration with AI as a screening tool.

Publisher

Springer Science and Business Media LLC

Reference33 articles.

1. Fecher, B., Hebing, M., Laufer, M., Pohle, J. & Sofsky, F. Friend or foe? Exploring the implications of large language models on the science system. AI & Soc (2023).

2. Single-reviewer abstract screening missed 13 percent of relevant studies: a crowd-based, randomized controlled trial;Gartlehner G;Journal of Clinical Epidemiology,2020

3. An open-source machine learning framework for efficient and transparent systematic reviews;Schoot R;Nat Mach Intell,2021

4. The Development of a Living Knowledge System and Implications for Future Systematic Searching;Grbin L;Journal of the Australian Library and Information Association,2022

5. Are ideas getting harder to find? Am;Bloom N;Econ. Rev.,2020

同舟云学术

1.学者识别学者识别

2.学术分析学术分析

3.人才评估人才评估

"同舟云学术"是以全球学者为主线,采集、加工和组织学术论文而形成的新型学术文献查询和分析系统,可以对全球学者进行文献检索和人才价值评估。用户可以通过关注某些学科领域的顶尖人物而持续追踪该领域的学科进展和研究前沿。经过近期的数据扩容,当前同舟云学术共收录了国内外主流学术期刊6万余种,收集的期刊论文及会议论文总量共计约1.5亿篇,并以每天添加12000余篇中外论文的速度递增。我们也可以为用户提供个性化、定制化的学者数据。欢迎来电咨询!咨询电话:010-8811{复制后删除}0370

www.globalauthorid.com

TOP

Copyright © 2019-2024 北京同舟云网络信息技术有限公司
京公网安备11010802033243号  京ICP备18003416号-3