Skip to Main Content

Generative AI Guidance (Chat GPT, Google Gemini, Microsoft Copilot)

What is Generative Artificial Intelligence?

Generative Artificial Intelligence (AI) is a type of artificial intelligence that can create new content such as text, images, and code.  It does this by learning patterns from existing data, then using this knowledge to generate new and unique outputs. Generative AI can create highly realistic and complex content that mimics human creativity, and its capabilities are developing rapidly.

See the videos below from the LinkedIn Learning Course What is Generative AI by Pinar Seyhan Demirdag - AI Director at Seyhan Lee for more information on what is Generative Ai. 

 

The Importance of Generative AI

The importance of Generative AI video link.

How Generative AI is Different from Other Types of AI 

How Generative AI is Different from Other Types of AI video link.

Generative AI Tools

Generative Artificial Intelligence (AI) covers a wide range of tools and functionality. With any of these tools it is important to understand that the output received is based entirely on the information inputted. See our AI best practice guide for further information on creating prompts for AI. 

Here are a few examples of types of AI that are relevant to studying at university.

Text Generation

Tools such as ChatGPT (free*), Google Gemini (free) and Microsoft Copilot (free) are capable of generating text from user inputted prompts. This text is unique and is created in real time. If the same prompt was given to the same text based generative AI twice the result would written differently. 

Text based generative AI can be valuable for learning, as they can be used to reexplain concepts, to answer surface level questions, for initial idea generation and prompting, for seeing other perspectives, and proofreading. However with any information given you should still apply your critical thinking and problem solving skills in your learning. 

Text based AI sometimes generates inaccurate, misleading or biased information, and so you will need to use your research and critical thinking skills to ensure that information is correct. A particular area of inaccuracy is generated references. When asked for information to be referenced Chat GPT has been found to fabricate references and therefore should not be used for finding sources of information for assignments. Please see our searching for resources guide for support in research.

You should not use AI in a manner that constitutes academic misconduct. The Academic Regulations, section J2.1(iii) Academic Misconduct, Plagiarism, states that: "If a student's work is found to contain verbatim (or near verbatim) quotation from the work of others (including from Generative AI) without acknowledgement, then plagiarism has been committed".

This means that Generative AI should not be used to write your work. There are uses for AI such as Grammerly (free*) in proofreading, or in idea generation, but the content and research should be your own work. 

Image Generation and Editing

AI tools such as Microsoft Copilot (free)Padlet (free), DALLE 3 (paid - free through Microsoft Copilot), and Canva (free*) allow for written text prompts to be turned into images. 

Here is an example from the free image generator in Padlet from the prompt "A diverse group of university students play a board game in a beautiful library."

A diverse group of students play a board game in a beautiful library. Notice that on face value the image looks fine, however if zoomed in the people have unrealistic features

Have a look at the image closely. What do you see? Look at the individuals features. 

Images from free generation software often are okay at on a surface level, however they have problems replicating human features particularly hand and eyes. Paid for versions are often higher quality, but come at a price.

AI can also be used in image editing. Photoshop and Google's 'magic eraser' tools can be used for near instant AI assisted image editing.

Research AI

AI tools such as Consensus (free*) have applications in finding information. Users are prompted to ask Consensus research questions and are shown quotes from academic sources that may answer the question. Consensus has a synthesize tool that attempts to summarise the findings of papers that it can access. These findings are added into a Consensus Meter that highlights whether research agrees with the statement. Sources found in Consensus can be clicked on to see their abstract, how many times they have been cited and to view where you can access the full text.  

It is important to note that Consensus only summarises 10 papers at a time and so may be missing important information. As interpreting research is difficult Consensus recommends reading the papers that it has found and consulting additional information.  Furthermore the Consensus Meter claim to only be 90% accurate. 

Example: We asked Consensus "Can mindfulness improve sleep?"

Consensus found 16 "highly relevant results" and analysed 13 papers to create the consensus meter.

Consensus and other research based AI can provide an initial insight into an area, but should not replace your own research skills and abilities. It is recommended that you still find your own sources, and read any sources identified by consensus to check the accuracy of the AI's findings. 

* Whilst the tool is free there are paid for levels