AI and accountancy examinations: first steps in relevant skill development

Dr Shonagh Douglas, lecturer in accounting and finance at the Robert Gordon University, outlines ways in which we can take the first steps in developing examinable AI-related skills in accountancy subjects.

Accountancy education is heavily driven by examinations both in the professional and academic context. It might be easy to assume that AI is irrelevant to examination-based subjects as students cannot access AI during the assessment. However, as AI becomes increasingly widespread in the accountancy workplace, it is important accounting educators develop the relevant skills for using AI and AI output within examination-based subjects. 

This is undoubtedly easier said than done and seems, at times, like climbing a moving mountain. Here are some suggestions for simple ways to get students and tutors started using AI and for developing AI-related skills in examination-based subjects.

Developing skills related to computational AI output

Firstly, consider replacing some computations with a completed answer which could be tutor or AI-generated. This could take different forms.

The output could have errors, and students need to identify them and explain why they are wrong. This could be extended by requiring either an explanation of how to correct them or a communication, such as an explanation email, to a senior or the client. 

Alternatively, the output could be correct, and the students are asked to explain or analyse the information that has been given. 

Both could be used in many areas: financial journals, tax computations, investment appraisal, and budget preparation, to name but a few. While recognising preparation can help develop knowledge, moving some questions to this format can effectively shift the focus from memorisation to understanding and interpretation.

Preparing for narrative elements of examinations

AI also has several potential uses to help students prepare for narrative elements of examinations. This could include generating an essay plan, a summary of the key requirements of a standard or legislation or providing illustrative examples of a concept being discussed. The output generated must then be evaluated. 

You will need to guide students through this, which, depending on what you have used it for, the following checklist might help.

Essay plan

• Have all the key relevant areas from teaching been covered? If so, is this accurate and in line with teaching?  If not, what is missing? 
• Where has the knowledge been developed beyond what was covered in the teaching?  How can you check if this is accurate?
• Can you give each section in the plan a rough percentage weighting to indicate where words should be focused?

Summary of an aspect of an accounting standard

This applies equally well to legislation, guidance, or other technical material.

• Is this accurate?
• What evidence can support this?
• Is it missing anything?
• Can you apply this to a specific scenario?

Illustrative example

• Is the illustration relevant?
• How can you verify the accuracy of the example?
• Is anything missing from the explanation?
• Can you think of an alternative example — which is better, and why?

An aid to remembering

Many accountancy examinations require students to learn information and although the value of this is debatable, it remains a critical requirement in many examinations.  AI can provide a bit of fun and something more memorable to support this. Some examples follow. 

Write poems

This is one of my favourite uses of AI – it really can write a good poem! 

Students preparing these themselves, checking the output accuracy, then sharing in class can be a great memorable activity. I’ve included one on remembering the different types of depreciation below.

Depreciation Methods: A Simple Poem

Straight Line
In the world of numbers, clear and bright,
Straight Line Depreciation does it right.
From when we buy to when it’s done,
The cost spreads out, year by year, one by one.

Every year, the same amount,
Makes it easy to account.
Value drops, nice and steady,
Keeping things simple and always ready.

Reducing Balance
But there’s another way to see,
With Reducing Balance, quick and free.
Early on, the cost is high,
Then it drops as time goes by.

At first, the value falls fast and steep,
But slows down, a gradual leap.
This method shows a faster start,
Reflecting wear and tear’s true heart.

ChatGPT, July 2024

I’ve extended this activity and asked AI to write in the style of famous people with effective results.

Write a song

Ask students to write a song in the style of a particular artist — the rest of the class guesses who it is.

Again, students would need to check for accuracy before sharing. I’ve tried this out with good results — a song to remember your debits and credits in the style of Taylor Swift was really not bad but probably unlikely to make it onto her next album! 

Make acronyms or memorable phrases

Make acronyms or memorable phrases to remember rules or explanations. Sometimes, on this one, you have to go back and forth to get one you like, but this process can also help learning.

ChatGPT was less good at making up debit and credit jokes so I have left those out!

I hope you have found this useful and has given you a few ideas. It would be great to hear from anyone who uses these or indeed anyone with an interest in AI in accountancy education and would like to discuss this further.

About Shonagh Douglas

Dr Shonagh Douglas is an Accounting and Finance lecturer at the Robert Gordon University in Aberdeen. Shonagh has a keen interest in accounting education and educational technology.

Shonagh is the contributing editor for Accounting Cafe. You can contact her with suggested contributions.