Welcome to your guide to using Generative ArtificiaI Intelligence (GenAI) for research.
GenAI, such as language models and text generators, can be powerful tools for your research. ChatGPT, Google Gemini, and Pi are examples of these.
We are particularly referring to text-based tools, also known as Large Language Models (LLMs) or ChatBots, as they have potential for helping to navigate our information landscape and support research.
It's important to approach these tools with a critical and informed perspective. This guide provides some key considerations for using GenAI in your research.
Illustration: The brain connects - Illustration - Plugging into the Brain - Illustration: When brain activity will merge with computers / Photo © Mondolithic Studios/Novapix / Bridgeman Images
What is GenAI?
GenAI, or Generative AI, is artificial intelligence that generates text, images or other media. It uses pre-trained models to generate new material. The new material is generated in the moment of the prompt (your input) and is not reproducible.
Jisc and the National Centre for AI have produced A Generative AI Primer that is regularly updated if you'd like to explore further information about these technologies.
Please check with your tutors about what use of GenAI is acceptable in your work. This may vary according to the module requirements.
GenAI can help you generate ideas and get started with your research, but it shouldn't be your only source of information. Use it as a tool to supplement your research, not replace it.
You may want to use these tools to:
- Get a sense your topic area and interconnected ideas
- Come up with some keywords on your topic
- Develop search strings to improve your search for use in academic databases and library catalogues (Library Search)
- Suggest databases or journals that might be useful to explore for your topic
GenAI can create human-like text, but it's important to remember that it's still a machine learning model. It may generate biased, inaccurate, or misleading information. Always verify the information generated by GenAI with reliable sources.
If you use GenAI to help you discover sources, make sure to check the sources and ensure they are real and relevant to your topic. You can easily do this by searching Library Search, Google Scholar, or even a general web search on your preferred browser.
Our Lean Library browser extension can help connect you to Library resources when you’re searching the open web.
Responses may be limited by:
- The quality of the question (prompt) you provide
- The data it has access to – think bias, equality, inclusion, representation, as well as date range
- The reliability of content – these tools are prone to hallucinate (make up) e.g. resources that don’t exist!
Remember that all information is created by humans, and that GenAI tools use models that are created by humans. These tools are subject to the same issues as any other information source! Think critically about how you use them. One method that is good for thinking critically about any online information source is lateral reading.
Lateral reading is the alternative to just reading the page you're on (vertical reading). Lateral reading can help you think critically about the response you get from a GenAI tool, as it involves navigating away from the response to check its validity.
Try lateral reading by opening extra tabs in your browser to check a few things e.g.
- Are the references real?
- Do the authors, theorists, artists exist?
- Are there alternative points-of-view, factors, events that are not reflected in the response?
You might find this method helps you to reframe your prompt to improve the generated response. For example, you might have more information to direct the tool to give a more specific or more relevant response.
Interactions with GenAI tools might spark creative ideas for your practice. This shouldn’t be a surprise as creative ‘sparks’ can occur at any time through a whole variety of interactions with our environment, other people, as well as resources, media and tools. These sparks, or moments of inspiration, are part of the research experience in creative practice.
GenAI won’t do your creative work for you. The way you develop and shape your work from a spark will be your voice in your practice, which you need to own.
GenAI tools offer many opportunities for enhancing our interactions with information but use them wisely. Think critically about how you might employ them (prompts) and how you might use the responses (generated content).
Over-reliance on any resource or method of engaging with our information environment will ultimately be limiting. It is much healthier and more enriching to blend a variety of approaches and explore a range of different resources. Chance encounters with ideas and our improvisational skill in drawing these into relationships is still more powerful and creative!
So far, we’ve been referring to GenAI (LLMs/ ChatBots) as distinct from search tools but we’re already seeing this technology being used to enhance our search - from browser search engines to library catalogues.
Search engines, like Google, have always provided a different search experience from a library catalogue or database – one, broadly speaking, uses natural language and the other uses controlled language. Switching between the two search experiences can cause problems and frustrations! But essentially both require keywords to return the best results.
GenAI powered search tools may be able to accommodate more complex natural language but the way you frame your question or prompt will still be important, in fact it will be an increasingly valuable skill.
Using GenAI raises important ethical questions, such as plagiarism, copyright, and the spread of misinformation. Be mindful of these issues and use GenAI responsibly.
Here we highlight these ethical issues:
- bias
- transparency
- privacy
- copyright and plagiarism
- equality
- environmental impact
GenAI:
- can be biased if trained on biased data. This can have the implication of e.g. perpetuating stereotypes or responding discriminatively. [bias]
- is trained on large datasets, which may not be publicly available and therefore can’t be scrutinised. [transparency]
- has access to large datasets that may include personal information which may not be anonymised or encrypted. [privacy]
- could be trained on copyrighted works and potentially plagiarise these, which could infringe copyright law. [copyright and plagiarism]
- tools are not all available for free. Paid models have access to wider and more up-to-date data. [equality]
- technology is power hungry. [environmental impact]
Research is a journey through your understanding of your topic. It helps you explore, make connections and see possibilities – it is not about quickly getting to an answer. If GenAI technology can improve your experience of searching and help you uncover relevant, connected information then that’s great. Use with caution though – remember it can only use the data available and might not return relevant results.
For more information and guidance on using Generative AI in your research, please contact your Academic Liaison Librarian or attend our Bitesized workshops. Let's work together to use Generative AI in a responsible and informed way to enhance our research and learning.
Still not sure?
For more information and guidance on using Generative AI in your research, please contact your Academic Liaison Librarian.
Library Drop-ins
Short appointments for quick questions on using Library resources.
Book a Librarian
Book one-to-one research support from your subject librarian.
Research & Finding Information
Step by step guide to finding and evaluating information