Artificial intelligence and the future of the accountancy profession

Jun 02, 2020
The accountancy profession needs to engage with  how emerging technologies like artificial intelligence will disrupt traditional career pathways.

By Dr Patrick Buckley, Dr Elaine Doyle, and Ruth Gilligan

Information technology has become inextricably embedded in virtually every aspect of our professional and personal lives. Data about what we do, what we are interested in, with whom we communicate and where we go can all be captured and stored at a scale unimaginable even five years ago. Technology giants such as Google, Amazon and Alibaba are engaged in a competitive race to capture the data generated by this new reality, lending credence to The Economist’s claim in 2017 that “the world’s most valuable resource is no longer oil, but data”. The data captured is valuable for several reasons. For one, traditional activities such as advertising can be personalised and optimised to a revolutionary degree – think of Facebook. Data also allows companies to build entirely new products. For example, the utility of Google Search results depends on analysing what information others have found useful in the past.

A further value assigned to these data streams is linked to the development of artificial intelligence (AI). A host of mathematical and algorithmic tools – some novel, some more mature but turbocharged by the advent of big data – has propelled the development of AI. Leaving aside philosophical questions such as to what extent these systems are intelligent, every-day and now familiar examples of AI (Siri and Alexa, for example), are demonstrably practical and effective. These visible successes, combined with the breakneck pace of development, pose a multitude of questions about the impact of AI on our future – not least its impact on the future of work.

The future of work

Concerns about automation and jobless futures are not new. Two centuries ago, Ricardo proposed that technology caused unemployment. In the 1930s, Keynes predicted that new technologies would reduce the demand for human labour. In the 1980s, Leontief compared the role of a human in the modern economy to that of a horse in agricultural production – first diminished, and then eliminated by automation.

Until the advent of AI, the consensus was that such predictions were overly simplistic. While new technologies can have a destructive effect on a particular industry or sector, their introduction often leads to increased opportunities in other areas. The overall effect is to change the structure of the jobs market, rather than result in a reduction in the work available. The jobs eliminated by new technology are replaced by jobs requiring higher-order cognitive skills (e.g. a robot replaces a welder but requires a software engineer to program it). Though this can be frightening and stressful for individuals, at a societal level, as long as education and training enable people to adapt to changing conditions by acquiring new skills, the long-term impact of technological change on the jobs market should be positive.

The rise of AI has disrupted this consensus. In brief, the suggestion is that the human monopoly on tasks requiring significant cognitive processing is being broken. Education and training may become ladders to nowhere if AI systems that match or surpass human cognitive abilities are feasible. A glance at the world today demonstrates that many tasks humans once performed are being automated by AI systems, with virtually all studies showing that the process is accelerating as the capability of AI systems improves. For example, two Oxford economists, Frey and Osborne, predict that 47% of jobs in the US will be automated by 2030.

The impact of AI

Investigating how this disruption is likely to impact the accountancy profession, our research profiled the tasks that practitioners perform at different stages of their career and at three levels: trainee/junior, manager, and director/partner. We then calculated the probability of each task being automated by aggregating information from a range of sources, including academic studies and reports from professional, industry and government organisations.

Our analysis makes it clear that, taken as a whole, accountants perform an enormous variety of tasks for their clients and employers. Some tasks, such as preparing accounts or tax returns, are considered extremely vulnerable to automation. Others, such as designing effective financial control strategies for clients, building relationships, or mentoring juniors and trainees are not.

This feature of the profession has two implications:

  1. Given the enormous variety of tasks performed and roles fulfilled by accountants, assigning a single probability and suggesting that this represents an objective assessment of how vulnerable the profession as a whole is to automation is a simplification to the point of absurdity.
  2. The large number of tasks not vulnerable to automation means that for the foreseeable future, the profession as a whole does not face an existential threat. Tasks like designing effective tax strategies or the financial structures of businesses will require a mix of quantitative and soft skills as well as a deep, strategic understanding of the world beyond the capabilities of AI.

Career pathways

However, this does not mean that the profession can afford to be complacent. Analysing the potential effects of AI at different stages of a traditional career pathway reveals that the tasks vulnerable to automation belong predominately to early career stages. This is particularly the case for trainees/juniors, but also applies substantially to certain work at manager level. Therefore, while accountants may always be needed, the current economic case for most trainees and some managers may disappear.

This presents challenges for the profession. Most obvious is the need to redesign career pathways in response to these trends. A traditional career pathway through the profession follows the well-worn path of trainee to manager to director to partner. A key question for firms and the profession is how to replenish senior ranks if the bottom rungs of the career progression ladder are removed. If there are no trainees or junior staff, where does the next generation of managers, directors and partners come from?

A second, related issue is that of skills and knowledge development. Generally, the more experienced individuals in organisations perform the more cognitively demanding tasks. The tasks most vulnerable to AI automation are often seen as repetitive and undemanding. At first glance, the automation of such tasks may seem a positive development for employers and employees alike. However, this perspective takes no account of the knowledge and skills gained by performing these tasks in a real-world setting. For example, designing effective tax strategies requires experience that can only be acquired by spending time working on basic tax compliance. It may be possible to develop the skills and aptitudes required by more senior practitioners without a long, real-world apprenticeship. However, there is no evidence to support this position.

At the very least, it seems likely that the entry pathway to the profession will need restructuring, with substantial changes required to curricula and entry requirements. In an extreme case, firms may face severe skills shortages a few years after engaging in significant automation. Higher-order skills may atrophy and disappear due to the lack of entry-level positions rupturing the supply pipeline of employees capable of performing higher-order tasks.

Perception of the profession

A third potential issue is the attractiveness of the profession to new entrants. If some of the tasks traditionally performed by managers are automated, then this will presumably have the effect of reducing the total number of individuals required at this level. The profession may evolve towards a position where a relatively small number of individuals (say 5%) do high-value, well-remunerated work while the other 95% are relegated to low-value, poorly paid tasks. A rational and risk-weighing decision-maker, the very type of intellect the profession seeks to attract, may select away from careers where the odds seem stacked against being able to access opportunity. In the long run, this selection bias may have a significant adverse effect on the profession’s ability to attract high-calibre candidates.

The future of the profession

Forecasting the future is a notoriously uncertain endeavour. Any predictions regarding the impact of AI on the accountancy profession (including those in this article) should be treated with scepticism. Reports of the imminent demise of the accountancy profession are, in all likelihood, greatly exaggerated. However, it would be equally short-sighted to discount the potential impact of AI on the profession entirely. It does seem likely that in the medium-term, the traditional career pathways associated with accountancy will be significantly dislocated. Responding to this will require meaningful, profession-wide dialogue and debate about how the next generation of accountants will be recruited, educated, and motivated.
 
Dr Patrick Buckley and Dr Elaine Doyle lecture at the Kemmy Business School, University of Limerick, and Ruth Gilligan is a Tax Associate at PwC Ireland.