Is AI coming for your job?

You’d be forgiven for anticipating a robotic replacement for your job when you hear the term ‘AI’, but the future of accountancy is decidedly human according to the Institute of Financial Accountants (IFA). We asked IFA Technical Manager Matt Barton for his insight.

Should we be afraid of AI, Matt?

In my opinion, no. AI presents an incredible opportunity for accountants and for the sector, and the savvy among us will be upskilling as soon as possible. Personally, I think the fear comes a little bit from the world of science fiction, where we’ve all read or watched stories about computers taking over from their human creators. There’s also the general fear of the unknown, which is understandable but should be approached critically. We’re a long way from a robotic uprising in the accountancy sector, or indeed in the world, so it seems unlikely it is something we should be afraid of. In fact, the opposite.

Will AI be taking over our jobs?

Sort of, but not in a scary, science fiction way. AI is an incredibly diverse field of study. There are a lot of individual disciplines, looking at lots of different ways we can streamline and improve our jobs and job functions. When we think of autonomous robots or computer systems, we are in fact considering a sub-division of AI called “neuromorphic”, which refers to AI that has the ability to mimic a human brain. In theory, it can ‘learn’ and deliver an agile response to inputs, making it autonomous from its creator. This type of AI is not really able to run programmes, or deliver repeatable functions. It is also an extremely complex and expensive discipline of AI, meaning that its future applications will be focused on maximum gains, for example in healthcare, rather than accountancy. When I discuss anyone’s fear over job security, this neuromorphic AI is usually the kind of threat they’re envisioning. I doubt that is a reality we will see in the majority of our lifetimes, or that we will be a high-priority sector for neuromorphic AI.

At the other end of the spectrum is algorithmic AI, which focuses on AI executing a repeatable function, guided by a set of appropriate rules. We are all familiar with algorithms, even if we don’t know we are, utilising them in everything from our social media networks to search engines, GPS applications to suggestions on what to buy or what to watch. They all utilise predictive algorithms to make suggestions or present information, based on how we have interacted before. It is this area of AI that will have the fundamental impact on our job functions, and I for one am excited about the prospect of an industry where the more basic, repeatable functions are streamlined, giving us more time to share and utilise our core expertise.

So, what does the future of accountancy look like?

There are two distinct areas that will be impacted by AI, and indeed which are already being impacted. The first is in our day-to-day job functions, the tools we use, and the way that we run our businesses. AI, or perhaps better called automation, is sweeping through most commercial functions. For example, Microsoft is one of the key investors in ChatGPT, and is therefore adopting its capabilities into Microsoft365 software. This will see it deliver all sorts of new features, such as predictive suggestions of responses to emails, to analysis and summary of customer interactions, and even task scheduling and meeting transcribing. It will basically automate and predict some of the most repetitive and boring tasks, freeing your time to focus on your business and your expertise, and less on your administration.

The second will be accounting-specific functions, working through the cloud-based accountancy software that has seen widespread adoption. These softwares already have the capability to deliver and streamline many of the mundane accountancy tasks, including data entry practices, and complex reporting. In fact, Making Tax Digital and iXBRL filing are both driving much of the demand for this improved automation. For some, this feels threatening, but I feel it presents an unprecedented opportunity to respond with client-first practices that benefit them, and benefit your own business too. It also has the power to boost productivity and therefore margins, and to help tackle the skills shortage.

The other significant advantage of digitisation is the opportunity to better interrogate data. Digital systems are typically cloud-based, offering real-time insights into business performance that can be shared between accountant and client. Accountants can embrace this recently accessible pool of data by offering client-focused services including data analysis, integrated reporting and flexible business modelling. It is about becoming a genuine strategic partner to clients, not just doing their books, but helping drive the business forward.

Are there any disadvantages to AI?

Absolutely. The biggest single disadvantage of AI is that they are perceived as infallible, but are in fact fallible. Coming back to neuromorphic AI, where many of us have a mistaken expectation that AI can ‘think’ like we can, but is potentially ‘better’ at it, or at least better at spotting mistakes. The reality is that AI is both the result of human thinking, and used by humans, who often have a divergent approach to that of an algorithmic AI. As the National Cyber Security Centre highlights in their advisory on large language models (LLM) like ChatGPT and Bard, they are definitely fallible. As LLM are reliant on the data that they receive, they can get things wrong, they can ‘hallucinate’ incorrect facts, they can be biased, they can be gullible, and their responses are directly related not just to the question that is being asked, but the way that question is posed. In terms of application, this gives it significant restrictions for current purposes, particularly in a field such as accountancy which requires at least some black and white thinking. What’s more, these more futuristic applications will require just as much training for the human as they will for the AI.

More: Chat GPT: UK accountants tussle over adoption timeline

Any final thoughts?

There is one additional aspect that will determine the adoption of AI within the accountancy sector, and that is our regulatory requirements. We are a compliance-heavy sector with a fiduciary responsibility that means we cannot automatically switch to an automated practice. In the same way that AI can offer benefits in overcoming human error and data mistakes, humans will, at least for the foreseeable future, need to oversee the application of AI to ensure that it is robust, and meeting our regulatory needs. For the hesitant among us, this is good news in that it will slow the pace of initial adoption, giving us all the time to catch up.

In conclusion, I think that future success will be dependent on our ability to innovate. These tools present significant opportunity, and the savvy among us will implement continuous improvement practices to ensure we are trained, ready, and willing to use the latest technology to our advantage. I for one am excited.