LibGuides: Research Skills: Using AI tools for research

AI for information searching

Unfortunately, popular free generative AI tools like ChatGPT and Google Gemini are not reliable search engines. They work by making predictions based on data they have been trained on, rather than retrieving live information.

Do not use them for finding information, because you will have to fact-check everything, which is a lot of work.

However, there are reliable AI tools that can help with certain types of research projects, including literature and systematic reviews. Read more below.

For internet searching, use Microsoft Copilot or Perplexity AI, but be sure to click on the citations so that you use information directly from the original source.

AI for academic research

AI search tools do not rely solely on traditional keyword searching, they use algorithms to find related articles.

Many of the well-known AI literature review tools, like Elicit, ResearchRabbit, and Consensus utilise another AI tool called Semantic Scholar. It has a unique feature of using TL;DR (‘too long didn’t read’) summaries instead of abstracts. (Read more about TL;DR summaries https://www.tldrthis.com/.)

One of the advantages Semantic Scholar offers is that it accesses large open-source repositories like Open Alex, PubMed, Springer Nature, Taylor & Francis, SAGE, Wiley, ACM, IEEE, arXiv, and Unpaywall.

Traditional Approach to conducting academic research

· Step 1 - Search your own institution's library collections with carefully selected keywords

· Step 2 - Identify the important journals associated with your field of study within your institution's collections, and look through the last three years of published and peer-reviewed articles.

AI-assisted approach

You still need the first two steps – AI does not replace your discipline-centric human judgement; it is a tool that complements your subject knowledge and allows you time-saving innovations.

Steps 1 and 2 will allow you to establish credible journals to find good seed articles to train an AI algorithm that works for your subject. Training them on a poorly researched or off-topic article will cause your search to take you further away from good articles.

Below is a summary of AI research tools for academic research purposes. You can find out more by watching the videos on the right-hand side of the page.

ResearchRabbit	Searches for academic literature Learns from your selections to find relevant literature Allows you build research collections Notifies you of recent research on your topic/s Provides interactive visualisations of topics and authors to expand or narrow your research Is collaborative (you can share your collections with others)
Lateral	Searches across multiple documents at once Searches contextually rather than relying on being fed the right keywords Makes suggestions for appropriate content based on selections made by the user Is collaborative (you can share your collections with others)
Elicit	Searches for academic literature Uses semantic search, so will encompass relevant keywords without you having to input them Summarizes research Finds relevant literature through searching for meaning and not just relying on accurate keywords Keeps refining results according to relevance that you determine Limited free access: paid plan required after a period of use
Open Knowledge Maps	The visual maps provide an instant overview of a topic by showing the main areas at a glance, and documents related to each area. Clusters similar documents together. Labels research areas with relevant concepts. Highlights open access content
Consensus	Search engine that ranks academic literature based on relevance, quality and accuracy Gives the consensus of the scientific community about your topic Lists the papers used to provide the consensus and includes links to full text versions
EvidenceHunt	Uses AI to search for clinical evidence Use simple search terms, predefined medical specialties, or use your own custom PubMed query Subscribe to a weekly e-alert. EvidenceHunt will run your search for you on a weekly basis so you get the most up to date clinical information Extracts summaries and sample sizes
Semantic Scholar	Searches research papers across all fields of science Papers can be saved and organised into a library, which Semantic Scholar learns from and make recommendations about other papers Sends out notifications of new research

Awesome AI tools for research

ResearchRabbit

Science Grad School Coach. (2022, November 1). How to Create a Literature Review Outline with Research Rabbit || Research Rabbit Tutorial [Video]. YouTube. https://www.youtube.com/watch?v=S2ygJv4lpWo

Lateral

Lateral. (2021, May 28). Welcome to Lateral! [Video]. YouTube. https://www.youtube.com/watch?v=9J2ZTNEgAg0

Elicit

BU Cares (2023, January 28). How to use AI for research: Elicit.org for writing a literature review [Video]. YouTube. https://www.youtube.com/watch?v=Oy2myE76ZfY&t=2s

Consensus

Consensus. (2022, September 7). Consensus - Product Hunt Beta Launch Demo [Video]. YouTube. https://www.youtube.com/watch?v=2_xxFEO8Hws&t=2s

Semantic Scholar

Semantic Scholar. (2022, March 16). Semantic Scholar: Product demo [Video]. YouTube. https://www.youtube.com/watch?v=U25ZhuokyG4

Open Knowledge Maps

Gray Man Concepts. (2021, March 11). How to use Open Knowledge Maps (open source intelligence, OSINT tools) [Video]. YouTube. https://www.youtube.com/watch?v=xIiw0GlJ5T0