TAKEAWAYS
“Are we going to be entirely replaced by AI in the future?” It is a question that we hear a lot from many aspiring accountants and auditors. With artificial intelligence (AI) advancing at unprecedented speed and news of AI breakthroughs flooding the news media, it is easy to fear that machines will soon be doing our jobs. Powerful generative AI models like GPT, which can write fluently and handle complex tasks, have further intensified these worries.
But is the prevailing narrative of AI replacing auditors actually rooted in reality? As researchers who have spent the last five years studying AI’s impact on auditing, we have uncovered some surprising truths that challenge this notion in a research study recently accepted by Management Science. Our data analysis, based on over 400,000 auditor resumes across the US, suggests that AI is working alongside them as a helpful co-pilot rather than replacing them outright.
To understand how AI is truly shaping auditing, we analysed hiring patterns and job postings at audit offices across the US. We tracked the hiring of employees with specialised AI skills, such as machine learning programming and natural language processing, to identify when each audit office began using AI. We then examined how the introduction of AI affected overall auditor headcount, audit quality, and fees.
Contrary to fears of job displacement, we found that audit offices that adopted AI actually increased their auditor headcount compared to non-adopting offices. This growth was concentrated among junior and mid-level auditors. Why? AI enables audit teams to take on more clients by automating routine tasks, freeing human auditors to focus on higher-level work like analysis and client interaction. In other words, AI can increase the volume of clients that auditors serve, which explains why more auditors are needed.
However, the auditor role is evolving. Job postings at AI-adopting offices prioritised soft and cognitive skills like critical thinking, efficiency, communication, and client service over technical AI expertise. This evidence suggests AI is handling the repetitive number-crunching as a co-pilot, allowing auditors to focus on work requiring professional judgement and people skills. In an AI-enabled audit, human skills are becoming more important, not less.
Crucially, we also found that the use of AI improves audit quality. AI-audited clients had more accurate assessments of going-concern risks and internal control issues. They were less likely to require serious financial restatements. By flagging high-risk areas and anomalies, AI helps auditors zero in on matters requiring professional scepticism. Auditors who were relieved of mundane work by AI can devote more time to complex issues demanding their insights.
To gain insights into how AI is being implemented on the ground, we also conducted indepth interviews with experienced audit partners from major US audit firms.
The partners shared that AI adoption typically follows a top-down model, with major investments and tool development happening at the national level, while individual offices have discretion in how they deploy AI based on their needs and readiness. Audit partners described how they have local champions who lead AI initiatives and support teams in using the AI tools effectively.
However, the partners also highlighted challenges in AI implementation, noting that some audit offices were hesitant to adopt it because of the learning curve involved. Interviewed partners admitted that there is always some fear and resistance to change, as not everyone has the technical skills or comfort level right away. Partners emphasised the importance of training, pilot testing, and gaining buy-in from audit teams to smooth the transition.
Clients’ IT systems and data readiness were another common hurdle. Some clients, especially smaller ones, are not set up to provide data in a format that AI can easily digest. Auditors sometimes have to work with clients to get their data house in order first. This shows that AI success in auditing depends not just on the audit firm’s capabilities but also on the technology preparedness of their clients.
For companies being audited, the interviewed partners believed AI would ultimately lead to a better experience and more valuable insights. By automating routine tasks, they can spend more time understanding the clients’ business models and providing proactive advice. It is a chance for them to be better partners, not just gatekeepers. Another added that with AI, they can analyse full data sets instead of just samples, meaning they can give clients a more holistic view of their financial health and risks.
The implications extend to investors and the broader market as well. AI-powered audits should give investors greater confidence in financial reporting and could help build higher trust. However, partners recognised the need for clear regulatory standards and governance around AI use, with one partner noting that they will need guidelines on things like AI model validation and controlling for machine bias. Explainability of AI algorithm and ethics will also be key to maintaining audit integrity.
Our findings paint a hopeful picture: AI is acting as a co-pilot rather than wholesale replacing auditors, boosting audit efficiency and audit quality while driving demand for uniquely human skills. Auditor judgement remains indispensable even in an AI-driven environment. However, as the technology matures, it will undoubtedly transform how audits are done and what skills are valued.
Looking further ahead, it is hard to predict exactly how AI will reshape auditing in the long run. With tools like GPT demonstrating the immense potential of artificial language, AI could conceivably handle much more complex audit tasks involving reasoning and judgement in the future. The auditor’s role may evolve into a more interpretive one, focused on interpreting AI-generated insights for clients and stakeholders. After all, auditing is not just a technical skill but a social endeavour built on human trust and accountability in the capital market.
For auditors-to-be, AI’s rise presents both challenge and opportunity. Tomorrow’s auditors may spend less time on routine tasks and more time on problem-solving and liaising with clients. Soft skills and critical thinking will be crucial. Auditors who can translate AI insights into real-world business guidance will have strong career prospects. Some baseline AI literacy will also be a must – auditors should be comfortable working with AI tools and interpreting the outputs. Auditors who embrace technological change and cultivate their human comparative advantage can look forward to a smarter, more impactful career. The most successful auditors will be those who can use AI insights, professional wisdom, and people skills to deliver top-notch audits.
The implications are clear for accounting educators and policymakers too. Accounting curricula must evolve to build AI-fluent, soft-skilled auditors. Regulators will need to set robust standards around responsible use of AI in auditing, from validating AI models to safeguarding independence and mitigating machine bias. Collaboration between academia, industry, and government will be key to getting this right.
In the end, while the AI age may stoke some existential angst, our research suggests AI will remain auditors’ co-pilot, not competitor, in the near future. By automating mundane tasks, AI will allow auditors to focus on the profoundly human skills that make their profession vital. The road ahead is one of upskilling and adaptation, not obsolescence.
AI is here to stay in auditing and will become an indispensable co-pilot for navigating an increasingly complex audit and accounting landscape. But human judgement, creativity, and trust will always be needed in the demanding work of keeping our financial system sound. Aspiring auditors and accountants should focus on sharpening their critical thinking and technological literacy. AI may be a gamechanger, but it is still a tool in human hands, at least for now. The full paper can be found here.
Kelvin K.F. Law is Associate Professor, Nanyang Business School, Division of Accounting, Nanyang Technological University. Michael Shen is Assistant Professor of Accounting, NUS Business School, National University of Singapore.
© 2024 Institute of Singapore Chartered Accountants