Unlocking the Potential of Artificial Intelligence in Accounting Audits

In a recent trial, Ernest &Young (EY), a prominent Big Four accounting firm, employed an artificial intelligence (AI) system designed to detect fraud in the financial accounts of select UK audit clients. The preliminary results were impressive, with the system uncovering suspicious activity in two out of the first ten companies examined. Subsequent investigations confirmed both cases as instances of fraud, highlighting the significant promise AI holds for enhancing audit quality and streamlining workloads.

The Potential of Artificial Intelligence

Kath Barrow, EY’s UK and Ireland assurance managing partner, emphasised the groundbreaking potential of AI-powered systems in analysing vast datasets. Proponents argue that these tools could become invaluable in alerting auditors to signs of wrongdoing and other issues.

However, the adoption of AI in auditing is met with varying opinions within the industry. Some audit firms express skepticism, questioning the ability of AI systems to reliably detect diverse forms of fraud with limited high-quality information. Privacy concerns also arise when confidential client data is utilised to develop AI algorithms.

Divergent approaches are evident among the UK’s major audit firms. While EY remains optimistic about the technology’s application for auditing, Simon Stephens, AI lead for audit and assurance at Deloitte, emphasises the uniqueness of fraud cases. He contends that AI, currently, may not effectively identify the novel patterns associated with each fraud.

Regulatory bodies, such as the UK’s Financial Reporting Council, recognise the potential of AI in enhancing audit quality and efficiency. Jason Bradley, head of assurance technology, has urged firms to possess the necessary expertise to ensure AI systems meet the required standards.

Traditional audit software relies on predefined patterns to identify fraud, whereas AI systems leverage machine learning and data from past cases. EY’s experiment utilised a machine-learning tool trained on numerous fraud schemes, both publicly available and from the firm’s past cases. The AI-assisted system, described as a “co-pilot” for auditors, focuses efforts on areas requiring further scrutiny.

Despite the optimism, some firms, including KPMG, question the capability of AI to detect sophisticated frauds due to the unpredictable nature of fraudulent activities. Deloitte currently limits AI use to less complex tasks, emphasising human oversight in critical areas.

Challenges persist, such as concerns about proprietary financial data and the need for auditors to understand the AI system’s coding and training data. Barrow acknowledges these challenges but underscores the increasing role of AI as a supplementary tool in risk assessment and identification.

The intersection of accounting and AI offers both promise and challenges. While AI has the potential to revolutionise audit practices, its widespread adoption hinges on addressing concerns, refining technology, and ensuring auditors are equipped with the necessary skills.

For further information or advice on your business, please contact us today!