Artificial Intelligence (AI) and Literature Searching

There are thousands of AI tools available, with a variety of specialities. Given that there are ethical and environmental issues with using many AI tools, we recommend considering the questions below in the Birmingham University Evaluative Framework before using an AI tool. Key points include:

• Do you need to use this tool? 
• Do you know how to use it appropriately? Are you allowed to use it for an assessment?
• Are you preventing yourself from developing research skills by using an AI tool for this task?
• Where has the information come from?

Don't forget to check your facts.

The University of Birmingham is developing an evaluative framework to help you self-evaluate an AI tool. 
We recommend considering the points below:

Relevance of tool

• Is an AI tool necessary for the task being undertaken?
• Has the tool been developed to perform the function you are intending to use it for?
• Are you preventing yourself from developing research skills by using an AI tool for this task?

Prompting and output

• Apply some practice prompts within the tool (before signing up, if possible) – what is the quality of the output like? Does it seem accurate and/or meaningful?
• Does the tool indicate if you are asking it to do something beyond its scope?
• Tip: use the TAP and TASTE methods for effective prompt engineering (Wallbank, 2023)

Policy compliance

• Does the way in which the AI tool will be used in your research fit the parameters of stakeholder policies relevant to you?
• Is your use of Generative AI in UKRI funding application process compliant with UKRI policy?
• Does the publisher you wish to publish with have a clear policy on the extent Generative AI can be used in your manuscript?
• Has your use of Generative AI been compatible with University guidance on using Generative AI tools ethically for work?

Corpus

• Does the tool make it clear what the corpus (the underlying data) is that is driving the AI?
• Is there a summary (single sentence/paragraph) of the methods used to develop the corpus and algorithm?
• Is it clear what the start and end dates are for the corpus data?
• Does the corpus raise concerns about built-in biases?

Costs

• Is the tool free to use or are there costs involved on a freemium or subscription basis?
• If the tool is available on a freemium basis, how much can be achieved before a cost is incurred?

Terms and conditions and data security

• Is it clear from the Terms and Conditions what the tool does with inputs and outputs it generates?
  • Who 'owns' that information?
  • Is it captured and reused?
• If you are using an AI Tool to analyse your research data, have you ensured you have ethical approval to do so?
  • Is using the tool part of your Data Management Plan?
  • Are there data protection issues?
• If the tool requires you to upload third party data, do you have copyright holder's permission to do so?

Reference: James Barnett & Lisa Bird (2025) Empowering AI literacy in the research community with an evaluative framework.