Skip to Main Content

AI | Artificial Intelligence: AI for Research

Artificial Intelligence and Research

 Artificial Intelligence (AI) is a tool, it is not a substitute for doing your own research. Though it has its strengths, it also has its limitations. Open GenAI tools scrape the internet for information. This may be good information, but it also could be misinformation and disinformation. Because it is software, it cannot tell if the information is accurate or not, even if it sounds convincing (Cottrell, 2024). You have to check the ethics, the accuracy and the references of any information that has been AI generated. If you use AI, you are responsible for any inaccuracies (O'Leary, 2024). As a researcher it is best practice to always critically think, analyse, reflect and check your sources.

Limitations of AI

  • It does not critically think or critically analyse information. 
  • It creates information that might seem okay and convincing but because it is machine generated and cannot 'think', the information can be inaccurate, bias or made up
  • Its only as good as the prompts you give it

Strengths of AI

  • Helps with time-consuming tasks.
  • Can reduce human error in data collection and analysis
  • Can help with brainstorming ideas and keywords
  • Can help with understanding or summarising complex topics

Evaluate What You See, Hear or Read 

Critically thinking, analysing and reading are essential aspects of tertiary education. Its important to question and reflect on all arguments, ideas and conclusions and not simply accept all that you see, hear or read (See the CRAAP test or Rauru Whakarare Evaluation Framework (a Kaupapa Māori-informed approach to evaluation) on the Finding Information and Research tab on your Subject guide). These skills are important for all information, including anything generated by AI.

Critical reading goes beyond just understanding the text. You need to analyse, interpret and evaluate the text i.e.

  • Do you understand the text and why it was written. Analyse the language, is it objective or subjective
  • Does the study make sense and does it do what it says it was going to do
  • Are there any strengths or weaknesses  i.e. are the arguments strong, consistent, convincing, balanced, and easy to understand or are there flaws in the author's reasoning
  • Are questionnaires, statistics or data gathered current and of an acceptable standard 
  • What are the techniques, strategies and evidence being used and does it correspond with what is being said in the conclusion
  • Are there any hidden agendas or bias (what organisations are the authors affiliated with)
  • Does it contribute to the field of study. Look at other experts in the field and compare, does it reach a similar conclusion or a different one. Why?
  • Is it accurate and based on reliable and credible research. Check the references. Are they scholarly and peer-reviewed

Self-reflect

  • How do you respond to the text
  • What are your biases, are you keeping an open mind
  • Does it answer your questions, is there anything the author isn't saying. 

(Cottrell, 2024; Duncan & O'Connor, 2004).

References: See Books/eBooks on this page.

Library Subject Guides

Please go to your library subject guides and check out the Book/eBooks and Databases tabs. Also, see the Finding Information and Research tabs and the Research guide for information on how to conduct effective research. If you need further help, please contact your Reference & Liaison Librarian.

Postgraduates and Research Papers

Check with your supervisor or the author guidelines of publishers. There may be advice on AI use when conducting research or writing books and journal articles i.e.

Reputable AI Research Tools

Different programmes may have different positions on the use of AI. Check with your kaiako/tutor if they have any guidelines on its use before starting your research

As a researcher, you have the subject expertise in your field. AI does not replace this. Below are some reputable AI research tools that can find open access research. Usually, they cannot access academic sources found in library databases because they are behind a paywall. Therefore, use these AI research tools along side library databases to get the best results.

Semantic Scholar

  • This is a free AI research tool for scientific literature
  • It is easy to use
  • It is regularly updated and checked (unlike Google Scholar) and peer-reviewed
  • It is mostly open access. If there is no full text available, search for the title in our library
  • It has a cited by feature, a related papers feature and a PDF reader
  • You can create an account to save resources, create alerts and export articles to referencing software such as Zotero

 Keenious

  • This is an open large language model tool powered by OpenAlex that helps researchers access knowledge and information that they may not have known existed
  • It has free and paid content (subscription-based)
  • It has a cited by feature
  • You can break it down into topics and locate similar articles and themes
  • You can import and manage your references in a similar way to referencing tools such as Zotero 
  • You can bookmark information

Dimensions

  • It has free and paid content (subscription-based)
  • It accesses open access publications on subject areas such as nursing, social work, sustainability etc
  • It has grants, patents, clinical trials, policy documents, data sets and reports on any research topic
  • It is easy to use. It has an advanced search, filters (i.e. publication year, type, title etc) and you can save items to a favorites folder
  • It can be used with referencing software such as Zotero 
  • It can summarise and give key highlights of a study and what top keywords have been used
  • It has a cited by feature

Core

  • It provides a comprehensive bibliographic database of open access scholarly literature delivered by the Open University (UK)
  • These are collections of research outputs from worldwide repositories and journals
  • It contains over 316 million articles 
  • It locates similar articles you can cross reference

Research Rabbit

  • It is an online citation-based literature mapping tool
  • It scans for any publicly available scholarly sources and selects articles based on their similarities
  • You can create a repository which is shareable
  • It is free

Lateral IO

  • Helps to read, find, share and organise research
  • Has a privacy policy, where you can upload documents and only you can see them in your workspace
  • It has free and paid content

(See also Wintec AI Literacy Toolbox: AI for research)

AI and Māori Data Sovereignty

AI & Māori Intellectual Property. Information on the AI & Copyright tab in the Copyright guide

Indigidata Aotearoa. Growing generations of Indigenous data scientists to protect Indigenous data so communities can be empowered by its possibilities.

Te Mana Raraunga/Data Sovereignty - Information on the Te Ao Māori guide

AI Pacific Data Sovereignty

Moana Connect. (2024, November 26). Moana talks: AI Pacific data sovereignity [Video]. YouTube. https://www.youtube.com/watch?v=sF14JF8n3OA

AI Frameworks and Guides (New Zealand)

AI Tool Quick Guide

Note. AI tool quickguide. Adapted from Academic Integrity: Using AI Tools for Assignments by Wintec, 2024 (https://libguides.wintec.ac.nz/academic-integrity/using-ai-tools). Reprinted with permission.

Journal Articles

A selection of articles on AI and Research. Copy and paste the article title into the search box on the library homepage. Click the down arrow next to the search box to select the Search Everything filter. 

Butson, R., & Spronken-Smith, R. (2024). AI and its implications for research in higher education: A critical dialogue. Higher Education Research & Development, 43(3), 563-577. https://doi.org/10.1080/07294360.2023.2280200

Choi, E. P. H., Lee, J. J., Ho, M., Kwok, J. Y. Y., & Lok, K. Y. W. (2023). Chatting or cheating? the impacts of ChatGPT and other artificial intelligence language models on nurse education. Nurse Education Today, 125, Article105796. https://doi.org/10.1016/j.nedt.2023.105796

Cotton, D. R. E., Cotton, P. A., & Shipway, J. R. (2024). Chatting and cheating: Ensuring academic integrity in the era of ChatGPT. Innovations in Education and Teaching International, 61(2), 228-239. https://doi.org/10.1080/14703297.2023.2190148

Flanagin, A., Pirracchio, R., Khera, R., Berkwits, M., Hswen, Y., & Bibbins-Domingo, K. (2024). Reporting use of AI in research and scholarly publication—JAMA network guidance. JAMA: Journal of the American Medical Association, 331(13), 1096-1098. https://doi.org/10.1001/jama.2024.3471

Kovoor, J. G., Smallbone, H., Jenkins, A., Stretton, B., Santhosh, S., Jacobsen, J. H. W., Gupta, A. K., Zaka, A., Nann, S. D., Jiang, M., Luo, Y., Withers, C., Ataie, S., Nematzadeh, N., Warren, L. R., Marshall-Webb, M., Chan, W., McNeil, K., Gluck, S., . . . Bacchi, S. (2024). The future is bright: Artificial intelligence for trainee medical officers in Australia and New Zealand. Internal Medicine Journal, 54(11),  1909-1912. https://doi.org/10.1111/imj.16518

Luo, X., Chen, F., Zhu, D., Wang, L., Wang, Z., Liu, H., Lyu, M., Wang, Y., Wang, Q., & Chen, Y. (2024). Potential roles of large language models in the production of systematic reviews and meta-analyses. Journal of Medical Internet Research26(11), Article e56780. https://doi.org/10.2196/56780

Vaughn, J., Ford, S. H., Scott, M., Jones, C., & Lewinski, A. (2024). Enhancing healthcare education: Leveraging ChatGPT for innovative simulation scenarios. Clinical Simulation in Nursing, 87, Article 101487. https://doi.org/10.1016/j.ecns.2023.101487

Check list for References

You cannot rely on GenAI to create original work or have information that is accurate. ChatGPT is not designed as a search tool. Its function is to generate and create information. Due to this, it has been known to make up references that do not exist. This is known as hallucination.

Source: Adapted with permission from Wintec [Academic Integrity: using AI tools for Assignments].

Therefore you should check your references. 

Checklist for References

  • Where did the reference come from? A library database, a reference list, Google/Google Scholar, ChatGPT?
  • Author: Who wrote it? What are their qualifications? What institution are they affiliated with? Look at the author's research output. Have they written other books, articles or carried out other research on this topic? 
  • Date: When was the information written? Does the date align with the research output of the author. Does the date match the title?
  • Title: Look at the title. Though it might exist, was it written by this author and at that date. Go to the library homepage, copy and paste the title in the search box and select the Search Everything filter. If it is an article, select scholarly & peer-reviewed on the left under Refine your search. If the library has access to it, it should come up. 
  • Source:
    • Journals: Search for the journal title in W & W E-Journals or locate the journal's website. Search within the journal by article title or by digging down via the year, volume, issue, page numbers. Can you locate the article title in the journal that the reference says its in?
    • Book/eBook's publisher: Check the library catalogue to see if the book/ebook is there. If not, go to the publisher's website and see if you can locate the book title.
    • A DOI is a string of letters, numbers and /or punctuation that identifies content and provides a reliable, persistent link to its location on the internet. If a DOI is available, click on it. Does it work or does it take you to an error page? If it does work, are the details the same as your reference? i.e. author, date, title, source.

Some library databases are now incorporating AI tools which act as research assistants. If you have used AI tools to help produce information, create a paraphrase, or you have quoted from ChatGPT, AI-powered research assistants or any other software, you need to include an in-text citation and reference list entry. If you are citing an AI summary based on a document, you should also cite the document the AI summary is based on. See AI and Referencing for more information.

Copilot

Copilot for Microsoft 365 is an AI assistant created by Microsoft to work with Microsoft 365 apps. It has been developed to help search for specific information, create and summarise text and create images based on text prompts. The technology is based on machine learning and is similar to ChatGPT. The free version of copilot accesses the internet but not MS Office 365. The paid version, Copilot for Microsoft 365 gets its information from the internet and is connected to the Microsoft graph so all the data in your MS Office 365 suite (organisational data) e.g. Office documents, One Drive, emails, Teams, SharePoint etc. Its knowledge is everything that you as a logged in user has access to and is inside your apps. See here for more information. (NB. It has a disclaimer stating you need to check for mistakes).

Therefore:

  • Treat copilot in the same way as ChatGPT
  • If you use copilot for finding information, you need to check all information for inaccuracies. Click on any references provided to go to the original source
  • There are concerns about privacy and data security. Copilot for Microsoft 365 states it does not retain prompts or train the large language model, but if permissions in MS 365 aren't carefully restricted, it can inadvertently result in access to confidential files, personal information, and intellectual property.  
  • There are concerns about bias

How to remove the Copilot for Microsoft 365 icon from Microsoft Edge:

  • Open the Microsoft Edge web browser and click on the main 3-dots menu button present at the right-side of the toolbar
  • Select the Settings option from the main menu
  • Click the Sidebar tab or section present in the left-side pane of the Settings page
  • Under the App and notification settings section, click the Copilot option
  • Turn off the "Show Copilot" toggle switch to remove the button from the top-right on Microsoft Edge
  • (Optional) Turn off the "Always show sidebar" toggle switch to remove the sidebar on the right

Please Note: You cannot remove the feature from the browser as it's built in, but you can hide all mention of it.

Tips

  • Keep draft copies of your research and your research process. You may need to show that what you have written is your own work.
  • Be transparent. You may be asked to declare the AI tools you have used and how   
  • Do not upload journal articles or published content that comes from library databases, ebook collections, online journals, standards etc to AI. This is a breach of copyright as the library has agreed to terms and conditions of use for many of its e-resources. See the AI and Copyright tab.
  • Do not upload your personal information or any personal information of research participants to AI. See the Ethics and Privacy tab
  • It is important that you are an honest and responsible scholar / researcher and acknowledge the work of other authors you have directly or indirectly quoted. Therefore, make sure you reference correctly and obtain information from a reputable source
  • As a researcher, you need to provide evidence that backs up your argument. It demonstrates your ability to sift through and read a variety of information, make sense of it, then link it to your writing. 
  • You must give your reader the ability to find original sources in case they wish to use them themselves or check the quotes you have used to make sure they are up to date, reliable and a true account of the original research or study. See the Checklist for References box
  • As a researcher you need to demonstrate your ability to research widely and thoroughly. You do this by searching for information then referencing it correctly. This shows that the information has not been made up, the meaning changed or misunderstood. 
  • You need to be careful with references created by ChatGPT as it has been known to create information and make up references that do not exist. Referencing information correctly protects you from allegations of plagiarism. See the Checklist for References box and the Plagiarism box on the APA Referencing guide

Books/Ebooks

Duncan, J., & O'Connor, M. (2004). Reading critically. University of Toronto Scarborough. https://www.stetson.edu/other/writing-program/media/CRITICAL%20READING.pdf 

Referencing Software

You can use referencing software such as Mendeley and Zotero to help collate and reference your work. They can produce an AI count in Turnitin, so talk to your tutor or supervisor about declaring their use. Also remember these references are machine generated so may not always be correct.  Check the APA Referencing guide before submitting any work.

Videos

O'Leary, Z. (2024, October, 22). Artificial intelligence and its impact on the conduct of research [Webinar]. Sage.

Library Services

If you have any questions, don't hesitate to contact the library team!

You can:

call 0800 141 121

email LSS@wandw.ac.nz

or pop into one of our campus libraries