The age of AI and considerations for Internal Audit

We are already into 2023, and based on all the reports and articles that were flooding the internet just a few weeks prior to the close of 2022, I am quite excited at the opportunities that lie ahead for the various professional disciplines in 2023, and in specific for the Internal Audit profession!

Yes – I am referring to the new AI tools in the market – ChatGPT and furthermore as outlined below that are already in existence, or on its way!

ChatSonic an advanced AI chatbot helps with real-time data, images, and voice searches, can also create content requirements based on user input, and generate content including the latest data and current trends. Other AI tools include LaMDA, developed by Google; Jasper Chat launched by Jasper; Chinchilla by Deepmind and so forth. Each of these AI tools offer a range of features that are advanced.

A few more inputs on the latest sensation – ChatGPT. Chat GPT, which is a Natural Language Processing (NLP) model, stands for?Chat Generative Pre-Trained Transformer. This is a chatbot based on the GPT-3.5 language model developed by the artificial intelligence research group OpenAI, and launched in November 2022, and employs deep learning to produce text that resembles that of a human where users can ask questions, and the platform uses artificial intelligence (AI) to provide replies. Users can receive both technical and non-jargon-based responses, in a natural way, much like a personal tutor, who is well-versed in all subjects.

What are the features of ChatGPT?

Key features include:

a) Responding to questions and answers

b) Composing texts, such as basic academic articles, literary texts, movie script, etc.

c) Text summarization and keyword detection

d) Fix and debug (For example, detect and correct errors in software code block)

e) Translation amongst various languages

f) Solving math equations, and many more…

The other tools mentioned above also offer a range of capabilities such as creation of content based on user input, perform internet searches and with the help of machine learning and NLP, it provides insights into what’s trending, ongoing news, and so forth.

Now neither is the purpose of this post to examine the pros and cons of each AI tool, nor does it purport to showcase the technical features of the various tools available.

The primary purpose is to bring to attention the fact that AI is a stark reality and is here to stay, and also highlight the various possibilities that can be leveraged by IA and perhaps the risks that go hand in hand with the advent of such advancements in technology. As a Risk management and Internal Audit (IA) professional, the questions that come to my mind are the following (arranged in no specific order):

a)?????How will the approaches to provide assurance change by IA profession?

b)?????How can the Internal audit function audit AI? What technical skills and subject matter expertise should they possess for identifying the right risks relating to such tools?

c)??????Furthermore, what is the role to be essayed by the IA function when an organisation is considering implementation of such AI tools or alternatively during such implementation?

d)?????How does an IA professional view with the right amount of professional skepticism the results generated by the Machine learning software? In specific, it is well known that biases do creep in the machine learning software and they do provide harmful results. A few known examples are the biases reflected in hiring by a leading retail organisation, or generating racist results by a healthcare network.

e)??????Will Internal Auditors be diligent enough to undertake independent and objective examination of data and ascertain facts before corroborating with AI generated results?

f)???????Where will the accountability vest if there any potential financial / legal ramifications arising out of decisions driven by the responses / outcomes of these bots?

g)?????As an Internal Auditor, would you blindly rely upon the recommendations or responses provided by AI tools such as ChatGPT?

h)?????Will the IA staffing budgets undergo significant reduction?

i)???????Would these AI tools put many of the IA & risk assurance professionals out of vogue?

j)???????How extensive have been the testing performed on the algorithms before the same has been adopted or implemented in the organisation?

k)?????What will be the impact of tools such as ChatGPT on report writing skills, which is where the output of IA efforts culminates?

As far as the possibilities go as enumerated above, AI tools will definitely enhance productivity, by eliminating many of the mundane / repetitive tasks that consume time.

But a point to ponder upon – Will AI totally replace human ingenuity in decision making?

The saying is “Time will tell”!

Look forward to views from my learned friends!