Leveraging the Potential of Generative AI to Accelerate Systematic Literature Reviews: An Example in the Area of Educational Technology

Abstract

Generative Artificial Intelligence (AI) is dramatically changing the way people work in many industries, including academia. Beyond its use for teaching, generative AI can also have a major impact on accelerating research processes. For example, generative AI can facilitate the identification of relevant articles when conducting a systematic literature review (SLR). This article compares six AIs (Forefront, GetGPT, ThebAI, Claude, Bard, and H2O) with their respective large language models (LLMs) when classifying 596 articles in the screening phase of an SLR. This SLR is aimed at exploring the development of non-technical skills with the support of technology in the field of medical education. Forefront with the LLM GPT-4 was the AI that obtained better results. The impact of this research is expected to contribute towards automating some of the phases of SRLs. Nevertheless, it is important to keep in mind limitations associated with the technology used to support this research, such as the rapid changes that AIs and their LLMs are currently undergoing, or potential restrictions on the number of requests per minute that AIs can receive as well as restrictions on the geographical location (since not all these AIs are available in all countries).

Ähnliche Arbeiten