What is generative AI?
Generative AI refers to tools that create new content such as text, images, video, audio, code or synthetic data. They use sophisticated machine learning algorithms that learn from vast amounts of data to produce new content that is similar to what they've learned. Essentially, these algorithms analyse existing content, such as text or images, and then use that information to generate new content in response to requests, known as prompts, from people.
The chatbot style interfaces used by these technologies are designed to simulate a human conversation, with text responses that appear human-like because they were trained on content written by people. However, this capability can sometimes blur the lines between human and machine interactions, potentially leading to misunderstandings or over-reliance on automated systems. It’s important to remain aware that, despite their sophistication, these tools lack genuine understanding and should be used with care.
View all Close all
Strengths of Generative AI
- Providing responses to questions or requests for information in a range of formats.
- Supporting the development of ideas or helping with problem-solving.
- Effectively performing analytical tasks and providing insights into data.
- Summarising information in a more digestible format.
- Continuously improving their capabilities through regular updates.
Limitations of Generative AI
Generative AI tools have several limitations due to their lack of understanding of the content they produce which can lead to misinformation, bias and an absence of ethical judgement. The content generated is based on predicting the next plausible word or sentence and they are trained to give responses that people like rather than responses that are necessarily accurate. This can result in:
- Inaccuracies: The generated content might not be factually correct. For AI-generated computer code, this could mean security issues, bugs or illegal use of software libraries.
- Biased and offensive content: The content may contain biases (e.g. political and cultural) and potentially offensive material.
- Fake references: The content may contain plausible, but entirely made-up references or citations.
- Dataset limitations: Generative AI tools are limited by the dataset they have been trained on. This may mean it might not include curriculum-relevant information and will not have access to scholarly databases, journals, and publications.
It is important that you understand the limitations and use your developing professional judgement to check the factual accuracy and appropriateness of the content produced. AI-generated content should be used together with the learning resources provided by your module tutor and other trustworthy sources.
Use of generative AI is generally permitted to support your learning; however, use of generative AI for assessments needs to be different to ensure that you have reached important learning outcomes. For assessments, use of generative AI is only permitted, where you have been instructed by your module tutor or programme director to do so. You can find details of how you can use AI for both your learning and assessments below.
Generative AI may be used to support your learning. You may find it useful to use generative AI to clarify key concepts or understanding, or to ask you questions based on your revision notes. To ensure the outputs from generative AI are useful, you will need to give it sufficient detail to support your request, for example including the level you are studying and the context. You must ensure that you review carefully the response you receive for accuracy, reliability and credibility.
We expect students to use generative AI tools in an ethical and responsible way. One aspect of this is that you must be aware of the limitations of these tools, in terms of reliability and bias, and ensure that you use generative AI to support your learning, rather than do your learning for you. Do not use generative AI tools as a replacement for the resources supplied by your module tutor. Learning to make effective, critical and ethical use of generative AI tools is an important skill as part of preparing for your career.
Learning to use AI, understanding its strengths and weaknesses, may also become a useful employability skill and you may find some modules will embed the use of AI into assessments and other activities.
When it comes to using AI as part of an assessment, the University’s default position is that you must not use generative AI tools for assessment, unless you have been instructed to do so by your module tutor or programme director. Guidance about what is permitted will be provided as a clear written statement, which may be found in the assessment details, handbook or on the Virtual Learning Environment. It is important to follow this guidance provided adhere to the principles of academic integrity and misconduct.
If you are permitted to use generative AI for your assessment, you are expected to it responsibly and ethically. This includes being aware of its limitations, acknowledging the use of generative AI in your work, and using it to support your own original work.
The University requires that all work submitted for assessment is your own original work, or in the case of group work, the original work of group members. If your module leader has given you instructions on the use of AI for a particular assessment, you should follow these. In the absence of any such specific instructions, the key point is that when you hand in work for an assessment, you are stating that it is your own original work.
The purpose of any assessment is to measure what you have learnt and understood from the course being assessed. It is not possible to assess your understanding if you plagiarise, copy and paste text, use AI, paraphrasing software or essay mills. The University sees academic integrity as very important and takes academic misconduct very seriously, seeking at all times to rigorously protect its academic standards.
Under the Academic Integrity Procedure unauthorised use of generative AI tools could be considered:
-
Plagiarism – where text used in the assessment is the direct output from an AI tool. Plagiarised text which has been amended using further AI tools (e.g., paraphrasing) remains plagiarised. The sources at the end of this message give guidance on appropriate citation of external material.
-
Falsification or fabrication – where the outputs from using generative AI tools have produced sources or data that do not exist.
-
Contract-cheating – where generative AI tools have been used to produce the whole assessment.
Any students suspected of committing academic misconduct will be actively investigated and if the misconduct is confirmed the full sanctions outlined within our Academic Integrity Procedure will apply.
Students should consult CiteThemRight for citing and referencing resources correctly. If the output of the AI produced material is available online, you can cite this as you normally would. If not, then you are advised to cite this as ‘personal communication’. Refer to the guidance on citing generative AI from Cite Them Right Online (your University login is required).
Where you have been permitted to use generative AI in your assessment task, you may be asked to acknowledge how you have used generative AI, what you have used it for and to include a copy of the output generated, including relevant prompts. Please refer to any guidance from your module tutor or programme director about requirements for acknowledging use of AI.
It is your responsibility to ensure that the reader of your work can clearly distinguish between original content created by you and content generated or inspired by generative AI and other sources. If you use generative AI in your assessment in any way that breaches our Academic Integrity and Misconduct Policy, this may be considered to be academic misconduct. Academic misconduct is taken very seriously, and all suspected cases will be thoroughly investigated. As part of an academic misconduct investigation, you may be asked to disclose what generative AI tools you have used, how you have used them (including specific prompts), and you may be required to provide draft copies of your work or other relevant evidence. It is important to understand these Guidelines and use generative AI responsibly to avoid any violations of academic integrity.
City St George’s provides Microsoft Copilot free for all students and you are recommended to use this tool:
-
About Copilot: Copilot uses the latest AI models, including GPT-4, so offers similar capabilities to ChatGPT-4. It can generate text based on prompts that use text, files or images.
-
Accessing Copilot: You can access Microsoft Copilot at copilot.microsoft.com, through the iOS or Android app, or in the Microsoft Edge sidebar and Windows 11. Ensure you log in with your University credentials to keep your data secure.
-
Copilot privacy and security: Microsoft’s enterprise data protection is applied when you log in to Copilot with your University credentials, which means that data entered is kept private and is not used to train the underlying dataset. As with all generative AI tools, you should not enter personal, sensitive or confidential information nor submit copyrighted materials into the tool.
Find out more about Microsoft Copilot.
There are a range of other generative AI tools available. The most well-known is ChatGPT by OpenAI which uses a chatbot interface for users to engage with the system; however, there are other tools available, including those embedded in search engines, existing software (such as Grammarly or Office 365) and computer coding editors.
Most tools require you to create an account which may be free or come at a cost. Free accounts are usually limited in terms of functionality or the number of times you can use the tool. Some tools use the data you enter to train the generative AI large language model. It is therefore important that you familiarise yourself with the privacy policy for any tools you use and do not enter any personal, sensitive or confidential information. In addition, ensure that you do not submit any copyrighted materials into the tools.
To use generative AI, you need to provide a set of instructions, known as a prompt, for what you would like it to do. The prompt should describe what you want the tool to perform, details of the task and any specific requirements for the output, For example:
“I am a first year Economics student in a UK university and I need to understand the concept of supply and demand, can you help me learn more about this and find useful study resources?”
Having reviewed the output, you may need to provide follow-up questions or information to help refine the information received.
Note: It is important to use your professional judgement to check the factual accuracy and appropriateness of the content produced and to use AI content in conjunction with the learning resources provided by your module tutor and other trustworthy sources, such as the Library.
Support is available to assist you with academic skills and writing:
- Visit the Academic Success Centre to access one-to-one support with a range of study issues, including effective study strategies, writing academic assignments and English language help.
- Refer to the Assessment Support page for information on Generative AI: guidance for students, Guidance for Exams, and Past exam papers provision.
- Contact the Student Union Support page if you would like advice on responding to an Academic Misconduct allegation.