In this guest blog, Dr Joanna Young, Electv Training, discusses how AI tools can make the literature searching and reading process more efficient.
Do you ever feel like there is too much to read? That the mountain of digital or physical paper that is your reading list keeps growing and you haven’t even started to write anything yet? Over 5.2 million articles have been published so far in 2024 (according to the Dimensions database checked on 21.11.24) and this number continues to rise every year. If we can’t read faster, can we read smarter?
From sifting through countless research articles to structuring a coherent narrative, the academic writing process demands time, patience and considerable mental energy. The good news is that new AI tools can make the literature searching and reading process more efficient and this is not limited to the multitasking chatbots such as ChatGPT, Copilot and Claude (though these can be useful too). Many specialised AI tools designed to streamline aspects of the research process now exist and the following are three examples: Elicit, ResearchRabbit and ChatPDF.
Please note: while the University of Edinburgh recognises the value of staff exploring potential uses of GenAI and is committed to supporting staff through the provision of training and the exploration of innovative and responsible uses of GenAI, it is important to recognise that the use of external GenAI tools is often not protected and staff should be aware that sharing any data with these tools may compromise research integrity. Additionally, staff should be aware that some published works are copyrighted and that copyright does not extend to being shared with certain GenAI tools.
In all instances, it is recommended that staff use the internal data-secure GenAI tool ELM to ensure ethical use of AI tools. ELM is the University of Edinburgh’s secure access to large language models (LLMs), including the newest version of OpenAI’s ChatGPT, GPT-4.
In using GenAI tools for their work, staff are asked to adhere to the guiding principles on the following webpage: University of Edinburgh Generative AI Guidance for Staff.
Elicit: Find & compare relevant papers to get you started
Instead of manually searching for papers, Elicit enables you to input a topic or a research question then it generates a table of relevant articles retrieved from Semantic Scholar, arXiv and bioRxiv. Different columns in the table allow you to compare papers by research question, findings, methodology, research gap, conclusion and more. It’s a helpful way to quickly eyeball and compare a range of relevant papers before deciding which ones to read in more detail.
ResearchRabbit: Visualise connections between papers
Articles reference more articles which lead to more articles: all researchers are familiar with “snowballing” when searching the literature. ResearchRabbit helps you to see how papers are connected through a simple interface and use of citation maps.
ChatPDF: Read papers quickly
In some fields, keeping up with the literature as quickly as it’s being published is becoming unmanageable. Yes, we will still read key papers of interest in detail, but when we want the key information rapidly then tools such as ChatPDF are very helpful. ChatPDF enables you to interact with PDFs as though you were having a conversation (with the AI). It can extract summaries and answer questions about the content – perhaps most importantly, you can sense check the outputs because ChatPDF links to the section of the PDF that it sourced the information from.
Using these tools can save you time and help in organising literature in a visually intuitive way. However, it’s essential to remember that AI can make mistakes or “hallucinate”, so do check the outputs.
We invite you to recommend more AI tools using the following hashtag on Bluesky #AI4Research Bluesky: joysci.bsky.social
Note: This post was written by the author, though ChatGPT was used to rephrase some sentences.
Electv Training (www.electv.net) delivers the Generative AI & the researcher: strategies, insights & practical uses workshop for researchers at universities around the UK & Europe. There is also a bitesize on-demand version of this course here: https://electv.teachable.com/courses/enrolled/2412287. The course will also be running as part of the IAD Researcher Development provision early 2025, further information can be found here: IAD Research Staff Course A-Z
The University of Edinburgh recognises the value of staff exploring potential uses of GenAI and is committed to supporting staff through the provision of training and the exploration of innovative and responsible uses of GenAI.: University of Edinburgh Generative AI Guidance for Staff