NARAYANAN VAIDYANATHAN, head of business insights, ACCA
Machine learning, machine intelligence, thinking machine, electronic brain – whatever you want to call it, artificial intelligence is here to stay. Although, machines haven’t completely taken over, they have slowly but surely crept into our lives affecting the way we live, communicate and ultimately work. From voice-driven assistance on a mobile phone, suggestive searches to autonomous driverless cars, we will continue to see fast-evolving technologies in the coming years.
At ACCA we have a deep interest in how technology impacts the accountancy profession and how it will continue to do so in the future. This year will see the 30th anniversary of the worldwide web – meaning we are firmly part of the digital revolution; technology is something accountants cannot shy away from or avoid.
So what is AI and machine learning?
The phrase artificial intelligence was coined 63 years ago by John McCarthy, an American computer scientist. A true artificial intelligence based system has the ability to learn on its own, understand information and make decisions in response to that information. In a sense, AI is the ability of machines to exhibit human-like capabilities in areas like thinking, understanding, reasoning, or perception. Machine learning is subset of AI that’s generally understood as the ability of the system to make predictions or decisions based on the analysis of a large historical dataset.
So why is AI relevant to the profession?
Embracing the capabilities of machine learning may not just be about competitive advantage; it could be about necessity. ACCA’s recent report, Machine learning: more science than fiction, takes readers through how new technological developments will have a huge impact on the accountancy profession. It is a subject we have to analyse to understand the value and benefits alongside the ethical challenges it presents. It is important for professional accountants to appreciate the value AI brings to the work they do – from generating valuable insights for business decision-making, to fraud detection, to risk assessment to understanding complexities in taxation and also with more effective non-financial reporting. So the accountancy profession needs to understand how AI and machine learning works, especially given its role in influencing the trust we have in the decisions of these systems.
Professional accountants also need to think about and mitigate potential risks from decisions made by an algorithm. With machine learning there is the risk of unintended consequences; for example, when bias in the input data used by an algorithm creates unfair or unacceptable outcomes. In this context, it will be important for professional accountants to develop a core understanding of these types of emerging technologies. The profession can truly benefit from the ability of these technologies to support them with intelligent analysis of vast amounts of data.
Can we trust AI and machine learning?
Machine learning is a powerful tool with tremendous potential, because it encompasses a staggering range of applications, such as recommendation search engines, fraud identification, detecting and predicting machine failure, optimising options-trading strategies, diagnosing health conditions, speech recognition and translation, enabling conversations with chat-bots, image recognition and classification, spam detection, predicting everything from how likely someone is to click on an advertisement, through to how many new patients a hospital will admit. While professional accountants may not need to develop machine learning algorithms themselves, they do need to have a sense of what they are dealing with. Accountants need to have the ability to ascertain whether decisions made by these systems can be trusted and viewed as ethical.
Are AI and machine learning a threat to the accountancy profession?
A recent Office for National Statistics survey, looking into professions at the highest risk of being automated, stated almost 26 per cent of jobs relating to professional accountants would be at risk from automation.
Certainly there is likely to be some disruption, for example in areas where the "intelligence" needed is in a specific, defined sense. An example might be the coding of invoices where the AI could learn from historic data and do the coding to high accuracy thereafter.
But even sophisticated technology like this, currently struggles to replicate the full contextual understanding and integrated thinking of which humans are capable. Therefore, human oversight of algorithm decision-making remains a consideration. And factors like building client relationships or leading successful teams remain deeply human endeavours.
Even beyond areas such as leadership, core technical activities can require forming an "opinion". This in turn involves judgement and interpretation that can draw on multiple factors, and where future choices may not be fully informed just by patterns seen in the past.
So, AI may not be ready to "take over" completely but the accountancy profession cannot be complacent and must prepare for the future.
With any use of technology, power comes alongside responsibility. In regards to machine learning, ethical considerations are paramount. Accountants need to consider and manage potential ethical compromise from decision-making by algorithm, such as the risk of bias in the data set that feeds them or the issue of who is ultimately accountable for decisions.
At the very least, all finance professionals should keep abreast of how AI is evolving and be wary of how the developing capabilities could overlap with their impact on their roles. To prepare for this, ACCA has already included a range of digital topics within its qualification. It has also enhanced the digital content across many of the exams for students, while also ensuring digital is weaved into members’ continuous professional development.