PCAOB explores generative AI

The Public Company Accounting Oversight Board issued a report outlining the results of its initial outreach efforts to auditors regarding generative artificial intelligence, representing early tentative steps by the board in evaluating how the technology is impacting audits.

The paper, released yesterday, reflects limited outreach to several audit firms and companies to understand their perspectives on the integration of generative AI in audits and financial reporting. This was part of the PCAOB's standard-setting agenda, which includes a research project to assess whether there is a need for guidance, changes to PCAOB standards, or other regulatory actions in light of the increased use of technology-based tools in the preparation and subsequent audit of financial statements.

The report says that, based on conversations with firm leaders, the use of generative AI seems to be focused primarily on administrative and research activities. Some firms stated that their staff can use generative AI when preparing certain administrative documents or initial drafts of memos and presentations related to the audit. Some firms also indicated that they had developed and deployed generative AI-enabled tools to assist staff in researching internal accounting and auditing guidance.

Meanwhile, firms also identified several potential areas where such tools may assist engagement teams in the future with planning and performing audits. The areas could include assisting with summarizing accounting policy and legal documents, evaluating the completeness of audit documentation against relevant documentation requirements, performing certain risk assessment procedures, scoping the audit, evaluating the completeness of financial statement disclosures, and comparing amounts in the financial statements or notes to the financial statements with audited amounts.

These findings are similar to those of a Gartner survey, which found only 12% of audit executives say they use generative AI for audit-specific use cases (see previous story). Of those who either are already using generative AI or plan to in the future, the primary stated use case was "Finding themes or patterns in large volumes of text," which was named by 65% of respondents. The second most frequently cited use case was in detecting anomalies or instances of noncompliance at 52%. Coming in third was drafting audit programs, with 50%. Meanwhile, 47% said they use it to summarize documents and 45% said they use it to draft audit reports. The least common use cases were writing code (18%), sentiment analysis and reviewing/correcting code (both 15%), scheduling meetings (11%) and, finally, determining audit ratings (8%).

The PCAOB report said that most audit firms interviewed noted the potential for using generative AI in certain aspects of planning and performing the audit. These firms indicated that PCAOB auditing standards are not currently viewed as impediments to the development and use of generative AI in the audit.

The PCAOB also observed that firms that are investing in generative AI-enabled tools indicated that they expect generative AI to augment, but not replace, humans in auditing or in financial reporting. In their view, human involvement remains essential for auditors and preparers and is needed to review the output from generative AI. PCAOB Chair Erica Williams expressed similar sentiments recently on her LinkedIn.

"Generative AI has significant potential to affect the way auditors plan and perform audits, and the PCAOB is watching carefully to ensure investors are protected. Ultimately, firms must have strong quality controls in place to ensure they are meeting their responsibilities to investors and the requirements under PCAOB rules and standards anytime they are considering using new tools, whether that is AI or anything else. Due professional care and exercising professional skepticism are essential when conducting an audit and cannot be replaced by AI," she wrote.

The PCAOB found similar dynamics at play with financial reporting: While some preparers are exploring opportunities to incorporate generative AI in accounting and financial reporting processes, integrating it in these processes appears to be secondary to integrating generative AI in operational and customer-facing areas.

This report represents one of the board's early forays into discussing the technology. The specific phrase "generative AI" has only recently been used in PCAOB documents, while use of the shorter "GenAI" is even newer still. The board, however, does have a robust history of discussing artificial intelligence in general.