Student and PGR guidance on using GenAI tools ethically for work

• Similar to how you might use the internet, Generative AI can help you understand different views, ideas, and perspectives. It can be a starting point for getting a handle on a topic area.
• Use it to define ‘task words’ in questions or assignment briefs so you are clear on what you should do e.g., defining the difference between ‘evaluate’ and ‘discuss’.

• The tools are very good at summarising information so you could use it to draw out patterns, themes, and topics areas from your notes.
• These summary notes could be condensed further into ‘flash cards’ useful for revision.
• Some tools can present a mind map of the information inputted. This could help with visualising your ideas or showing the relationship between themes.

• Use it to compare two answers given in response to a prompt question.
• Practise your evaluation skills by assessing the quality of arguments given.
• Check the accuracy of information generated.
• Find resources to back up or challenge the information generated.
• Use it to pose alternative arguments to claims you are making to test their validity and strength.

• Notes can be summarised into manageable chunks for revision purposes.
• Generate practice questions based on topic areas you are revising.

• The tools can help you break larger tasks down into smaller, manageable steps.
• Create to-do lists and identify objectives and tasks.
• Offer solutions and helpful tips to avoid procrastination.
• Help you devise a Gantt chart (a detailed schedule and timeframe) for longer projects.

• Provide a starting point for understanding different kinds of writing assignment e.g., exploring the differences between an essay, a report, or a blog post.
• Help you understand the key features of academic writing e.g., what constitutes an effective introduction or conclusion.

• It can act as a ‘third-party editorial assistant’, which means you can use the tools to correct spelling and punctuation, reduce sentence length to improve clarity, and ensure your writing is following correct conventions of grammar and syntax.  For PGR theses, this must be declared using the ‘third party editor declaration’ form.
• However Generative AI tools must not be used to alter your text in the following ways: to clarify and/or develop the ideas, arguments, and explanations; to correct the accuracy of the information contained within your work; to develop or change the ideas and arguments presented within your work; to translate your work into English; to reduce the length of your work.
• When using a Generative AI tool to correct grammar and syntax, you should not copy and paste large amounts of your own writing into the tool. You should apply the tool’s suggestions to your writing by paraphrasing or rewriting your sentences, rather than copying and pasting the tool’s output. See more information on privacy and copyright issues and what you should be aware of when adding your own text to these tools.

There are many tools available to help summarise information, make comparisons between text or data, and make reading articles easier and quicker. 

• However, there are potentially serious copyright implications for doing this and you should only upload other people’s work if the item is no longer in copyright or if you have permission to do so.
• Rather than supplying a full copy of a text by uploading or copying and pasting it directly into the Generative AI platforms, it is recommended that you point to the article title, DOI, or link for the Generative AI tool to then find the content itself and summarise it from there.
• See more information on copyright issues.

• Although tools like ChatGPT and Gemini respond to natural language (such as the language you might already use to search Google), using effective prompts can result in better answers from the tools. When we talk about prompts, we mean the questions you ask the tool to generate a response. There is a skill to asking the right questions so that the output you receive is relevant and useful.
• Try using an acronym like TAP to help you write an effective prompt. This considers the topic, output and parameters:
  • Topic: identify the topic area and make sure it’s focused, specific and given a context.
  • Action: be clear what output is required and for what purpose. Use action words such as ‘summarise’, ‘clarify’, ‘propose’. How do you want the information to be presented?  i.e., do you want an explanation or a comparison between ideas?
  • Parameters: this is where you fine-tune the response and give additional information to focus the results, like demographic, geographical location, inclusion of references, or date range.
• Once you have tried some prompts, you can then apply the acronym TASTE to improve and evaluate the content generated by the tool. This involves:
  • Test: Generative AI can be inconsistent with its responses so click ‘regenerate response’ to consider new perspectives.
  • Adjust: Vary and refine the prompt to compare the output.
  • Simplify: it’s very easy to provide the tool with a vague, ambiguous, or wordy prompt. If the output produced isn’t focused, try and simplify your prompt.
  • Trust: Examine the truth and accuracy of the output
  • Examine: Start to critique the outputs, focusing on the quality of the information and/or argument provided.

It is important to understand that the information provided by Generative AI may not always be reliable or up to date, for example they have been known to commonly ‘hallucinate’ references which do not exist.  Always double-check and verify information obtained from Generative AI tools. Cross-referencing with other reliable or peer reviewed sources is recommended:

• Use the Library Catalogue, FindIt@Bham, or an appropriate academic database. There are Library Guides which identify suitable resources for your subject area.

Apply critical thinking skills. Be critical of the text produced by Generative AI, just as you would with any other source. Ask questions of the information generated: 

• Does this information confirm knowledge I already have on the topic?
• Can the information be corroborated by outside sources?
• Does it include any bias, assumptions, or unverified opinions?

Whilst at University, you’ll hear about the importance of developing your Graduate Attributes. These are the skills, behaviours and values that you are developing to support your progress whilst at UoB and beyond. ‘Critical and creative thinkers’ is one of these Graduate Attributes and it includes the following capabilities: 

• To exercise discerning and autonomous thinking
• To think critically in order to reach a solution
• To analyse and evaluate the research you’re engaging with
• To handle and analyse data

These are skills Generative AI tools currently struggle to replicate. They can synthesise information but are less able to analyse and evaluate the information presented. Weighing up research and presenting a coherent argument that is backed-up with evidence is a very human skill and consequently you cannot rely on Generative AI to do this for you.

Your assessment is likely to be geared towards assessing your ability to think critically and creatively so it’s important that you focus on developing these skills. The Academic Skills Centre has an Introduction to critical thinking Canvas course that can help you hone this skill.

Generative AI is only able to generate responses based on the information it has been trained on and the datasets it has access to. Therefore, if the dataset has bias within it, it is likely to be transferred to the information generated. This means biases, stereotypes, and an over-reliance on a Western perspective can be passed on if careful analysis is not given to the information it produces. Therefore, do make sure you check the information generated against a wider variety of reputable and trusted sources.

You should think carefully about the consequences of pasting your written work into a Generative AI tool. It’s advisable to remove any identifiable information from work you submit into the tool.

It is necessary to ensure that any tools protect your privacy and personal data. There will be privacy notices to read which detail how the tools use any personal data you provide.  The UK Data Protection Act (DPA) 2018 sets out rules about how personal data can be collected, used, and stored. If you are using Generative AI tools to collect or process personal data, you need to comply with the DPA. This includes ensuring that the data is collected fairly and lawfully, and that it is only used for the purposes for which it was collected. If you are collecting or processing personal data using an AI tool, you need to be transparent with the individuals whose data you are collecting. This means informing them about the purpose of the data collection, and how the data will be used. If you include other people’s personal data into the tools an ethics review might be required.  

When using Generative AI tools, it is important to be aware of copyright laws as they will apply to the content uploaded to the Generative AI interface, the training data used to build the models, and potentially the outputs generated.

You should check the terms and conditions of any Generative AI tool that you use to understand the rights that you grant to the providers to re-use your work. For most Generative AI tools, by uploading your work you grant the providers the right to use it as they see fit, and this may include sharing it with other people. Your work will normally be incorporated into the training data and used to generate new outputs, at least in part. If you have really novel ideas or thoughts, you might not want to share these via Generative AI tools.

You should not upload content to Generative AI tools, or social media platforms, without the permission of the copyright holder. This includes text, images, data, songs, videos, and other copyright works, or any part of these. If the content is still in copyright the person that owns the copyright in the content is the only party that can grant the permission the tool asks for in the terms and conditions. If you do not have clear permission from the copyright owner to use their content with the Generative AI tools, do not upload it.

Please note that none of the print or electronic content held by the Library is currently licensed for such use. More about Copyright can be found on the Library Services' Copyright pages.   

It is important to be aware of the process by which generative AI tools are built and how these can present ethical issues such as social and environmental exploitation. For example, there are claims that developers have outsourced data labelling to low-wage workers in poor conditions.

Academic integrity is an important part of your university studies. This involves you completing your assessment honestly and respecting other people’s ideas. It’s about ensuring that what you submit for your assessment is your own work. The Code of Practice sets out the guidelines for academic integrity and your responsibilities as a student or PGR. As noted in Academic misconduct, the misuse of Generative AI tools in assessment, such as improper referencing or non-acknowledgement of use, will be dealt with in-line with this Code of Practice. It is also very important that you do not use these tools in assessment unless given explicit permission to do so. Therefore, to avoid plagiarism, it’s vital that you know when the tools can be used and how they should be acknowledged. Your School should provide you with this information, so please ask if you are unsure.