Skip to Main Content

Examples of use of AI in research

How AI can help researchers

A variety of AI tools can help streamline the process of discovering, managing and synthesising information. However it's important to note that use of these tools should be complementary to traditional research techniques such as structured, replicable database searching. Researchers will need to make an assessment of any tools used, and apply a combination of approaches, particularly in literature searching. For example, the use of AI tools alone is not supported by current best practice guidance for specific review types in certain disciplines (such as systematic reviews in health, where requirements are particularly strict due to real world implications for healthcare).

Literature search and discovery

  • Keyword analysis: AI can assist researchers in selecting and refining relevant keywords and search terms for their topic
  • Automated search: AI can perform automated searches across various databases, including academic journals, conference proceedings, and preprint archives
  • Google Scholar is an example of an AI-driven, free search engine for scholarly articles, theses, books, conference papers and patents. You can access Google Scholar through the Library catalogue

Literature review

  • Summarisation: AI can generate concise summaries of research papers, helping researchers quickly understand the key findings and contributions
  • Sentiment Analysis: AI can analyse the sentiment and tone of research papers to assess the general reception of certain ideas or concepts in the field
  • Clustering and Categorisation: AI can group related papers into thematic clusters, making it easier to navigate a large body of literature
  • Platforms like Mendeley and Zotero can help organise research and sharing with others

Citation and reference management

  • Citation Extraction: AI tools can extract and format citations from research papers, saving researchers time in managing references
  • Reference Management Software: AI-powered reference management tools help researchers organise and cite their sources efficiently
  • EndNote is an example of a popular AI-driven reference management software platform used for organising references and creating bibliographies
  • UniSA has an online guide to help with use of EndNote. Training workshops are also available through Edgex and myRD

Question answering

  • AI chatbots or question-answering systems can provide answers to specific questions about a topic by analysing the literature
  • Quillionz is an AI-powered question generation tool that be helpful for creating questions for study or assessments

Recommender systems

  • AI-driven recommendation systems can suggest relevant papers based on a researcher's interests and previous reading history
  • Collaborative filtering and content-based filtering can personalize recommendations
  • Elicit is an example of a tool designed to help with an entry-level exploration of scentific literature on a particular topic

Data extraction and analysis

  • Text Mining: AI can extract structured data and insights from unstructured text in research papers, enabling meta-analyses and systematic reviews
  • Data Visualisation: AI can create data visualisations, such as graphs or charts, to help researchers understand trends and patterns in the literature
  • Jupyter Notebooks is an open-source web application that allows you to create and share documents that contain live code, equations, visualisations, and narrative text
  • Excel is probably one of the most widely-used tools for data analysis

Real-time updates

  • AI can notify researchers of the latest papers, developments, or trends in their field of interest, keeping them up to date with the latest research
  • Google Alerts can monitor the web for interesting new content then send email alerts about its findings in real time

Collaboration and knowledge sharing

  • AI-driven platforms can facilitate collaboration among researchers by recommending potential collaborators based on their expertise and interests
  • Microsoft Teams and Slack are examples of collaboration platforms that are widely used

Plagiarism detection

  • AI tools can help researchers ensure their work is original by comparing it to existing literature and identifying potential instances of plagiarism
  • Turnitin and ithenticate are used at UniSA to detect instances of plagiarism

Language translation

  • AI-powered translation tools can assist researchers in accessing relevant literature in languages they may not be fluent in
  • Google Translate is a freely available example of a translation tool

Natural language understanding

  • AI-powered chatbots or virtual assistants can help researchers interact with scholarly databases and retrieve information more efficiently using natural language queries
  • Stanford NLP is a suite of NLP tools developed by the Stanford Natural Language Processing Group.

Using generative AI to develop research question

(video length: 3 min 45 sec)

Dr. Lynette Pretorius from Monash University shows how generative artificial intelligence can be used to develop one's starting research questions.

Examples

Example of using ChatGPT to help a PhD candidate generate research questions (taken from this journal article)

Use of AI tool can be openly reported eg. Responses to the following prompts were generated by ChatGPT (OpenAI, https://chat.openai.com/) on XXX date.

Core concerns within a research field could first be explored. Draft research questions could then be created to investigate some of the core concepts in a particular type of research study (eg. qualitative research study). AI could then help refine the question further with a focus on a specific sub-theme. A particular theory could even be applied and a new research question reformulated through that lens. Use of AI tools to finesse a research question would be particularly helpful for remote students who find it difficult to make time with a potential supervisor.

Researcher uses an AI tool to filter search results for a systematic review

The sheer volume of research papers and studies available can make it difficult to sift through the literature efficiently. AI tools can be utilised to develop a custom search algorithm that can filter and prioritise research papers based on relevance and quality. An example of one such tool that is available to UniSA staff is Covidence. Developers of that platform are introducing new innovations in automation, including the ability to sort by relevancy based on past screening decisions.

Researcher uses AI to generate complex code

AI tools can provide assistance to HDR candidates and researchers to write intricate, complex code saving significant time and effort. In the methodology section of the produced research, it is important to include detailed information about the tools used and how they were applied.

PhD candidate uses generative AI to write large portions of their thesis

There are ethical concerns about the use of generative AI in academic research, particularly when it becomes the primary author of a work. While AI can assist in various research tasks, it should complement the researcher's intellectual contribution rather than replace it entirely. Proper attribution and ethical guidelines must be followed in academic research.

Improving employability outcomes

When used astutely, AI could be employed in career exploration, informational interviewing and job application document writing, to help HDR candidates become more adept at promoting their value proposition to industry (high level research skills).