AI tools for researchers: Key insights for librarians to enhance academic support

Author: Saskia Hoving

Proactive librarians are key in helping researchers use AI tools practically and ethically, but both groups often find it hard to keep up with the many available options and their trustworthiness. AI tools like ChatGPT can change how researchers work, from literature searches to data analysis and publishing. Specialized AI tools help with specific tasks like forming hypotheses, analyzing data, and writing manuscripts. Librarians who provide the right tools make research more efficient and effective. This post looks at how AI impacts the different stages of research.

As gatekeepers and knowledge-bearers, research librarians are at the forefront of these tools. Understanding and advising on them is now a crucial part of their role, enhancing their value to researchers and promoting ethical and responsible AI use.

Building on the large-scale Nature survey released in 2023, we interviewed 13 academic researchers across APAC, Europe, Africa and North America, diving deeper into their experiences and views of using AI. This blog summarises these interviews, so librarians can get familiar with them and enable them to better support their researchers.

The mention of tools in this article does not imply endorsement; you use them under your responsibility. Also note that tools’ names, URLs, and other specifics are continually changing.

Where most people start: ChatGPT

ChatGPT is what’s called an AI chatbot. Its ability to generate seemingly coherent text in a simple user interface rocketed it to widespread use in early 2023. Users can interact with ChatGPT (and its competitors such as Google Gemini and Microsoft Copilot) in conversational text and many languages, allowing them to automate tasks such as finding and summarising information, editing text, translating, crunching data, and scanning and reporting on documents and websites.

AI chatbots are the AI gateway for researchers, librarians, and everyday users alike, and even seen as synonymous with “AI,” but there’s so much more. The array of existing and emerging tools is more focused and dedicated, meaning they can be more reliable and specifically helpful. They can be applied in the early, middle, and latter stages of research.

AI tools in the early stages of research

The early stages of research include formulating ideas and approaches, preliminary literature searches, finding funding and collaborators, and ensuring data sources' accuracy and reliability. These activities provide a foundation and roadmap for the research, to ensure that it builds on existing work while addressing new questions and lay the groundwork for its completion.

Literature searches
Researchers often find the early stages of research, especially literature searches, to be time-consuming and labour-intensive. Yet it’s essential to sift through the mass of literature to find relevant sources, ensure references show the latest findings, and synthesise information into a coherent literature review. "I find that the early stages of research present difficulties mostly related to research gap identification, innovation, and funding proposals,” says a PhD candidate at the University of Prishtina in Kosovo.

Generative AI tools can speed up and somewhat automate literature searches. A general chatbot can list areas to explore, and there are specialised tools that can present the relevant literature in accessible ways, summarise large volumes of text and complex ideas, extract key points, find commonalities/differences, and even extract data. “I’ve used AI tools, such as ChatGPT, Claude, and Consensus, to search for potential research gaps and ideas,” the PhD candidate says.

Experimental design
AI tools can also suggest potential sampling strategies and experimental designs based on the study's objectives and topics to help design research methodologies. A Librarian at Government Islamia Graduate College Civil Lines Lahore (Pakistan), notes that ChatGPT and Gemini can “help in formulating the research questions and objectives of the study, research methodology and research methods” and that they give researchers a “very good baseline.”

Finding collaborators
Finding suitable collaborators can be challenging because of geographical and institutional barriers. During the interviews, an African researcher noted, “Sometimes it is difficult for us in Africa to know what is going on in other places, where are people doing similar research or things we could collaborate on?”

There are an increasing number of AI tools designed to help researchers find connections between different studies and understand who is collaborating on a topic.

AI tools researchers are using in the early stages

• Assistant by scite – An AI-powered for literature retrieval, validating output against an internal database.
• Connected Papers – Visually maps connections between papers based on citation patterns.
• IRIS.AI – The product of an EU-funded project, it finds relevant papers, summarises findings, identifies connections between different studies, and extracts data. 
• Nature Navigator – Tracking of research trends, summarises articles, and identifies collaboration opportunities.
• Perplexity – Summarises complex documents, reducing reading time and making comprehending studies easier.
• Research Rabbit – Explores topics of interest by suggesting related papers, tracking associated research trends, and organising findings. 
• TDM – Automated process of selecting and analysing large amounts of text or data resources for purposes such as searching, finding patterns, discovering relationships, semantic analysis.

AI tools in the middle stages of research

The middle stages of research include trials, experiments, and data collection, along with data analysis and the actual writing up of the research manuscript. This phase is crucial for generating empirical evidence that supports or refutes the research hypotheses and preparing the work for its ultimate publication. Especially for writing, many tools have existed and been evolving since before ChatGPT. What’s changed is how they’re increasingly using machine learning and AI to enhance and expand their abilities.

Working with data
Organising, curating, and storing large datasets is, again, time-consuming and requires great attention to detail and a willingness to repeat and refine. Manipulating data between different software can lead to errors and inefficiencies. Researchers interviewed mentioned the inevitable burden of curating and storing data and welcomed tools that would help.

The PhD candidate at the University of Prishtina in Kosovo notes this stage can “be the longest and most difficult to deal with, [because of] sample selection, formulation of methodologies, and data processing”

AI tools can help automate data organisation—going beyond human-limited data analysis—and fix typos while improving tone and readability. They can also automate repetitive tasks and check data accuracy for trials, experiments, and data collection. AI tools can also provide advanced analytics capabilities so researchers can uncover insights they might miss in manual analysis.

Drafting papers
Writing papers, especially when English isn’t your first language, is a process many researchers would like to shortcut, and is also perhaps the greatest point of controversy in researchers’ use of AI tools.

A professor at the Armed Forces Institute of Rehabilitation Medicine in Pakistan list a number of tools he uses: “Writing a zero draft – Microsoft Dictate and AudioPen; improving the draft and creating and outline – Jenni and LLMs (ChatGPT, Claude, Gemini Pro, Mistral); writing the manuscript – Paperpal.”

However, researchers must be careful with the output, which often needs further refinement before it’s usable. AI tools should always be part of a broader analytical framework, combining AI-driven insights with human-ensured accuracy and relevance.

AI tools researchers are using in the middle stages

• AudioPen – Converts voice notes into text that's easy to read and ready to share.
• ChatGPT – Helps get answers, find inspiration and be more productive.
• Claude – Processes large amounts of information, brainstorms ideas, generates text and code.
• Dante – Assists with data analysis and visualisation, providing recommendations based on complex datasets.Gemini Pro – Recalls on long-context retrieval tasks across modalities, unlocking the ability to accurately process large-scale documents, thousands of lines of code, hours of audio, video.
• Gemini Pro – Recalls on long-context retrieval tasks across modalities, unlocking the ability to accurately process large-scale documents, thousands of lines of code, hours of audio, video.
• Grammarly – Well-known writing assistant that checks for grammar and readability.
• Jenni – Writing assistant to enhance academic and content writing.
• Microsoft Dictate – Uses speech-to-text to author content in Office with a microphone and reliable internet connection.
• Mistral – Provides a fine-tuning API through La Plateforme, making it easy to fine-tune our open-source and commercial models.
• Quillbot – Writing assistant for paraphrasing, summarising, and improving writing.
• Paperpal – A generative AI-powered academic writing tool.
• Research Round-Ups – Literature Reviews generated through a human-machine collaboration.
• Tableau – Business-oriented data visualisation tool for analysing and presenting data.

AI tools for the later stages of research

The final stages of research work are to ensure quality and integrity, find and submit to a suitable publication, and go through peer review and revision. These stages disseminate and validate research findings within the academic and broader communities.

Finding a journal
Finding a journal is a challenge and choosing the right one can dictate the time spent on revising, as well as the impact on the researchers’ publication record. Some experts advise selecting the target journal at the outset, but most researchers save that task for this stage. 

There are now many AI-assisted journal finders that can help researchers identify appropriate journals to submit their work.

Preparing manuscripts
Once a journal has been selected, preparing a manuscript for submission requires great attention to detail to meet the formatting, citation, and submission guidelines. 

AI tools can help researchers streamline these later-stage activities by automating many time-consuming processes in manuscript preparation and submission. For instance, AI can assist in formatting manuscripts according to journal guidelines, checking for proper citations, and ensuring adherence to style guides. AI-driven plagiarism detection tools can scan manuscripts to identify any unintentional similarities with existing work, thereby helping researchers maintain the originality of their submissions.

It’s important that researchers do not use these tools to cut corners. Researchers must consider that AI tools can yield formatting errors or miss field-specific, nuanced issues that a human reviewer may catch. Use AI tools as an aid rather than a replacement for human oversight, combining AI-driven insights with manual review to produce quality work.

AI tools researchers are using in the later stages

• Penelope.ai: Helps authors satisfy journal requirements.
• Journal finders: Jane (journal author name selector), Journal Selection dashboard (Dimensions with Altmetric data), and MY journal selector (Edanz), among many others
• Copyscape – Plagiarism detection tool that scans text for duplicate content and reports on potential matches. 
• Snapp - A peer review system that enhances the efficiency of the publishing process.

How librarians can support researchers in using AI tools

To help researchers use AI tools, it’s important that librarians experiment with these tools themselves, to collaborate with patrons and to see first-hand the output and the level of satisfaction with the results. Then, they can adjust their approach and offerings as necessary. It’s important to frequently revisit what’s on offer because this is a rapid and dynamic evolution.

“AI tools have made it easier to conduct research,” Waseem says, “but the researchers have to ensure the ethical use of AI tools in their work.” He recommends these sites for keeping up to date: AI News, Developer Tech, MIT News, and OpenAI News.

Tips for librarians
There’s no prescription for librarians on the “right” or “wrong” way to use AI tools or which to trust and recommend. These are choices you’ll need to make on your own through your professional experience and interactions. Here are some approaches for introducing AI tools to researchers:

• Demo workshops: Organise interactive workshops showcasing AI tools with hands-on sessions for researchers.
• Integration in research guides: Update research guides with step-by-step tutorials and case studies on AI tools.
• Personalised consultations: Offer one-on-one consultations to recommend and set up AI tools tailored to researchers’ needs.
• Use case competitions: Host competitions for innovative AI tool use proposals, with rewards like premium access or extra support.
• Awareness campaigns: You can run these using newsletters and social media, featuring success stories and a “Tool of the Month.”

AI tools are really changing the game for researchers, making everything from data analysis to writing up results a lot smoother. Librarians are key players in this shift, helping researchers get the most out of these new technologies while keeping things ethical. By staying on top of the latest AI developments, librarians can support researchers in navigating this exciting, yet complex, landscape. Together, they can make the most of AI’s potential, pushing the boundaries of what’s possible in academic research.

Curious about Springer Nature’s stance on AI? Explore our AI Innovation Framework, which aligns technological advancements with ethical and legal standards to benefit the research community.

Related content:

• What to know to plan research training on using AI
• AI in libraries: Improving tradition through innovation
• Making AI work for all of scholarly communications — sustainably 
• AI and Science Publishing: “Cutting through the clutter has never been more important”
• Can AI help us manage information overload?
• Unravelling the Power of TDM: A Transformative Journey for Researchers
• TDM and APIs: An introduction for information professionals
• Text and Data Mining: Uncovering Hidden Data Points and Powering New Discoveries
• The rise of generative AI: Transforming the role of information professionals
• Three ways Generative AI can give Research Managers “superpowers”

Don't miss the latest news & blogs, subscribe to The Link Alerts!