Education Officer Blog - 'The Robots Are Coming !' - A plain English guide to AI at the Uni of Bath


Blog Post

Education Officer Blog

'The Robots Are Coming !' - A plain English guide to AI at the Uni of Bath

If you're an Academic Rep, you have most definitely heard me wax lyrical about this already (and have it in your Useful Info channel on Teams!) - but AI is here, it is big, it is a little scary, and - oh. Wait. We have guides, lovely librarians, and integration into your assessment briefs. Lets talk about that. 

 

First and foremost, the guidance from the University across the Library, Centre for Learning and Teaching, and everyone else in that big ol' mix covers generative AI (or GenAI for short) - this is when AI makes something NEW for you - think ChatGPT summarising work for you or DALLE making a picture according to your prompts. Whenever I refer to AI - I mean GenAI. I am describing new terms throughout, but if you want a list you can just dart back to, I am popping this at the very end :) 

 

WHEN CAN I USE AI?

All assessments will have a type explciity listed in your assessment brief. This will be A, B, or C.

 

You can think of this like a traffic light - A is red, B is orange, and C is green.

TYPE A

Type A (or red) is when you cannot use AI in any way, shape or form. You can't, for example, use ChatGPT to give you ideas to start an essay. It is worth noting that plagarism checkers do not reliably check for AI, but you may still be challenged by your marker.

 

TYPE B

Type B,or orange, is when you can use AI as an assitive tool. The way you can use it will be explicitly defined in your assessment brief, like being able to use it to get you off a blank page by asking for prompts from ChatGPT. You don't have to use AI with this level, but if you do, remember to reference it. You can find guides on the University Library website.

TYPE C

Type C is a funky one - it is where AI has an integral role in your asessment. This, again, will be included in your assessment brief, and just as with type B, you must reference how and when you have used AI.

 

HOW DO I REFERENCE AI?

If you have a type B or C assessment and have used AI, then there are two general rules to how to reference your use of AI. The following all relates to Harvard Bath, so double check against what referencing style your course asks for.

 

If you used GenAI as something to assist your work (like asking for prompts)

Acknowledge it once with a general statement. The Library suggests something like:

"I acknowledge that this work is my own, and I used ChatGPT 3.5 (Open AI, https://chatgpt.com/) to summarise my initial notes and to proofread my final draft only."

 

The things you need to include are:

  • Name and version of the generative AI system used; e.g. ChatGPT-3.5
  • Publisher (company that made the AI system); e.g. OpenAI
  • URL of the AI system
  • Brief description (single sentence) of context in which the tool was used. 
  • Confirmation work is the student’s own.

 

If you quoted or paraphrased GenAI output

Use an in-text citation. The Library page is super comprehensive on this, so make sure to check that out, but in short this varies if it is text or photo, and if you can return to it via a link (retrievable) or not).

 

Retrievable text is referenced as a website/page.

 

Non retrievable text can be included in an appendix if you downloaded a copy, or as a 'personal communication' if you don't have a copy.

 

Retrievable images are referenced similiarly to web images, with a tool, year of publication, and listing the image reference to your reference list using the webpage format.

 

Non-retrievable images use an in text 'personal communications' citation in the caption. The Library gives the example of:

Figure 3. Shark in a library image generated using an AI tool (Craiyon, AI Image Generator (pers. comm.) 14 July 2022).

 

Once again, I highly encourage you to use this blog post as a kind of "oh thats not as scary as I thought" vibe for the referencing, and super encourage you to use the library guide and check in with your subject librarians if you have any doubts.

 

WHAT IF I USE IT WHEN I'M NOT MEANT TO?

If you use AI when you are not meant to for whatever reason, this sadly counts as academic misconduct (which is kinda like a fancy word for cheating).

As you would have done your academic integrity test, and ticked the box on moodle saying you agree to the Academic Integrity Statement when you submitted, it means you are acknowledging that you know when you can and can't use AI.

 

This can seem totally intimidating, and it might have been a while since you did your test, so it is worth checking if you're not sure, but in a nutshell, examples of misconduct can include:

Plagarism - using AI without citation, including paraphrasing

Collusion - Using information or answers from AI in a type A assessment, or without referencing in a type B/C.

Contract cheating (including impersonation) - presenting AI generated work as something you made, through purchase or any other means.

Breaching of examination regulations - going against the Academic Integrity Statement (this is kinda a catch all)

 

HOW DO THEY DETECT AI ?

Y'all I have seen the rumours, I have seen the panic, and I have to let you in on a little something something.

 

There are NO tools that can RELIABLY and ACCURATELY detect GenAI material. Not only this, they are prone to bias against certain groups, such as those who have English as an additional language.

 

So the Uni doesn't use them.

Instead, staff have a guide and relevant training for what they should do if they suspect you have used AI when you shouldn't have.

 

WHAT HAPPENS IF I AM ACCUSED OF USING AI ?

This is something at the forefront of the Education team's priority this year, and we even started collaborating with the Centre for Learning and Teaching early last year to ensure that the processes around these are fair, just, and simply not traumatic for you as a student.

 

It is something I am super passionate about, as I think NO academic processes should take up a massive chunk of a students time (we know you are busy enough !), and you should be supported throughout - no matter if you have actually conducted any kind of misconduct or not.

 

If you are accused of academic misconduct, you will be told and an investigation will be launched. This is where you can provide evidence to prove the work is your own. Even though this is an "investigation" it isn't that harsh (its all down to University lingo) - it is way closer to a discussion, and should be balanced at all times.

 

WHAT KIND OF EVIDENCE COULD I GIVE?

This is dependent on the kind of misconduct suspected, but could include something like:

  • showing a working reference list
  • document editing history
  • a viva where you get to talk about your process and findings

 

Okay my loves, if you have made it all the way to the end - great job ! I hope this demystifies everything AI for you even a little bit, and I'm wishing you the best of luck this year !

 

LoL (lots of love),

Amber

 

USEFUL TERMS
GenAI - GenAI is short for generative AI and refers to any kind of AI that makes something NEW.
ChatGPT - A conversation/chatbot style GenAI


Retrievable - Something that can be returned to via a link.
Non- retrievable - Something that can NOT be returned to via a link.


Academic misconduct - Basically a fancy word for cheating. It can often be interchanged with "breach of academic integrity" and "assessment offence" and includes stuff like plagiarism.
Plagarism - Claiming someone else's work is your own 
Collusion - Working together on an assessment where you are meant to work alone
Contract cheating - When someone else completes an assignment for you - this is usually for payment.
Breaching of examination regulations - Going against anything you agreed to in the academic integrity statement - colloquially known as cheating.


Centre for Learning and Teaching (CLT) - A department in the University whose job is to teach lecturers how to teach (in the absolute most simple terms) - they do work on everything from AI to feedback quality.