1. Artificial intelligence (AI) has the potential to transform literature reviews by expediting various steps of the process, such as problem formulation, literature search, screening for inclusion, quality assessment, data extraction, and data analysis.
2. AI-based tools can handle large volumes of documents and partially structured metadata, making them particularly useful for conducting literature reviews in information systems research and related social science disciplines.
3. The use of AI in literature reviews is still in its early stages, but there is a need for a comprehensive research agenda to explore the potential benefits and challenges associated with AI-based literature reviews (AILRs) and to shape their future development.
The article titled "Artificial intelligence and the conduct of literature reviews" discusses the potential use of AI in conducting literature reviews. While the article provides an overview of the topic and outlines a research agenda for AI-based literature reviews (AILRs), there are several areas where critical analysis is warranted.
One potential bias in the article is its focus on the benefits and opportunities of using AI in literature reviews, without adequately addressing potential risks or limitations. The authors emphasize how AI can expedite certain steps of the review process and reduce researchers' efforts, but they do not thoroughly discuss the potential drawbacks or challenges associated with relying on AI tools. For example, they briefly mention biases and model overfitting as challenges but do not delve into these issues or provide evidence to support their claims.
Furthermore, the article lacks a balanced discussion by not exploring counterarguments or alternative perspectives on the use of AI in literature reviews. While the authors acknowledge that there are two dominant narratives in the discourse on AI-based research practices, they primarily align themselves with the perspective that advocates for leveraging AI tools without critically examining potential concerns raised by skeptics.
The article also falls short in providing concrete evidence or examples to support its claims about the effectiveness of AI-based tools in literature reviews. While it mentions a few existing tools and studies that have used AI techniques, it does not provide sufficient evidence to demonstrate their efficacy or reliability. Without empirical evidence or comparative studies, it is difficult to assess whether these tools truly enhance the review process or if they introduce new biases or limitations.
Additionally, there is a lack of consideration for ethical considerations and potential risks associated with relying heavily on AI in literature reviews. The authors briefly mention black box predictions and acceptance by the research community as challenges but do not delve into broader ethical implications such as data privacy, algorithmic bias, or unintended consequences of automated decision-making.
Overall, while the article provides an introduction to the topic of using AI in literature reviews and outlines a research agenda, it lacks critical analysis, balanced discussion, and empirical evidence to support its claims. It would benefit from addressing potential biases, exploring counterarguments, providing more concrete examples, and considering ethical considerations and risks associated with AI-based literature reviews.