Skip to Main Content
ask us

WRHA Virtual Library: Artificial Intelligence

Artificial Intelligence

Artificial Intelligence (AI) are technological tools that are designed to mimic human learning processes and problem-solving capabilities. AI tools like ChatGPT are very popular, but using AI appropriately for research requires careful consideration.

Using AI for research

AI can be used as a tool for enhancing research. It can recommend keywords for search, brainstorm ideas and concepts, or summarize information. However, whenever using AI for research, be aware that it can "hallucinate" - for example, provide sources that do not exist. For more information on the uses and limitations of AI in searching, see this guide from Brown University.

AI tools for research include:

  • Research Rabbit: a literature review mapping tool
  • Consensus: a search engine that summarizes insights from research papers
  • Elicit: a research tool for summarizing and synthesizing papers
  • Scite: a tool for assessing how papers have been cited

Prompt design

Prompts are the text that you enter into an AI tool. A well-designed prompt can make a big difference in how effective AI will be in returning the sort of content you want. One framework you can use to improve prompt design is CLEAR:

  • Concise: Use clear language and avoid wordiness
  • Logic: Order your ideas in a way that makes sense and allows the AI to draw connections in the information you provide
  • Explicit: Include precise instructions to help frame your prompt to the desired context
  • Adaptive: If your prompt is initially unsuccessful, experiment with tweaks that might improve output
  • Reflective: Evaluate the response you get from the AI tool to understand how you can improve the prompt

Assessing AI output

AIs are not subject matter experts and may provide inaccurate information. Whenever using an AI tool, be sure to evaluate its output and check any sources provided. Cross-reference what the AI says with reference resources that are known to be reliable. Use AI as a starting point rather than an ending point of research.

The ROBOT test from McGill University can help you to assess AI tools.

Using AI for publication

Although AI tools are capable of outputting text, they cannot be accountable for the content of that text, so cannot meet the ICMJE criteria for authorship. They can be useful for brainstorming or creating an outline for a paper, or for providing writing assistance.

If you are using AI tools for something that you intend to publish, be aware that many publishers have imposed limits on how AI tools can be used. See this table from the University of Adelaide, but be sure to check with your specific publisher.

Citing AI

Any use of AI tools in publications should be acknowledged; be sure to indicate what tool you are using and how it was used. See this guide from Brown University Library.

AI and ethics

When using AI tools, it is important to be aware of ethical issues associated with them:

  • AI tools are susceptible to bias and can often perpetuate biases that are present in their training data.
  • AI tools may store and share data - it is essential not to enter any private information into an AI tool
  • AIs are often black boxes, which means it is not transparent where their information is coming from
  • AI tools can be used to replace human labour, for example in the creation of AI-generated artwork. They also scrape the work of artists and writers in developing their training set, and may as a result produce derivative works of copyrighted materials.
  • Large language model AIs like ChatGPT have a significant environmental impact.

To learn more, check out the UNESCO Ethics of Artificial Intelligence Hub or the University of Oxford Institute for Ethics in AI.