New Era for Auditing Profession Amidst AI Disruption
The auditing landscape is experiencing a seismic shift unprecedented since the corporate financial debacle over two decades ago that necessitated the formation of the Public Company Accounting Oversight Board (PCAOB) and precipitated a clear demarcation between consulting and auditing functions.
Unlike the previous crisis, today’s upheaval is not fueled by fraudulent activities or compromised auditor independence but rather by the meteoric ascent of artificial intelligence (AI).
This paradigm shift demands a thorough reassessment of auditing standards to adequately address the perils and prospects that AI introduces.
In a recent call for public comments, the PCAOB sought feedback on whether it should pursue standard-setting initiatives and investigations in this burgeoning area.
Ideally, this inquiry should have been merely rhetorical, as the affirmative response is unequivocally clear and should have commenced without delay.
The PCAOB has indeed made strides concerning AI integration. Last month, the board inaugurated an inspections modernization council aimed at enhancing audit quality.
Additionally, it consulted with firms on the application of generative AI in audits and has monitored AI’s utilization among auditors.
Former PCAOB chair Erica Williams underscored that the adoption of AI cannot supplant the auditor’s own expertise and discernment.
This initial foray represents merely a nascent effort, grappling predominantly with the application of legacy standards within a novel context.
It is imperative for the PCAOB to delineate unequivocal guidelines governing the acceptable use of AI to avert potential audit failures attributable to improper AI deployment.
One could reasonably conjecture that such a failure has already transpired, though it remains undetected.
Over the past two years, three of the four foremost global accounting firms retracted reports due to misleading AI outputs.
It is only a matter of time before an audit report is rescinded due to AI-generated errors or inadequate oversight of the AI mechanisms scrutinizing audit evidence.
Auditors are held accountable for the reports they endorse, regardless of the extent to which AI contributed—whether significantly, minimally, or not at all—paralleling their responsibility for the actions of staff, interns, and other team members. However, existing standards were not crafted with respect to the advanced capabilities enabled by AI.
Expecting auditors to adhere to current standards, many of which were conceived in an era when AI existed solely in the realm of speculative fiction, will inevitably engender inconsistent practices as each auditor and firm attempts to reinterpret antiquated rules for contemporary applications.
This scenario epitomizes the very purpose for which the PCAOB was established.
Auditing standards traditionally emphasize sampling and the interpretation of inaccuracies identified within samples. Only a few years ago, an auditor for a major public entity might inspect a mere fraction of annual revenue or expense transactions.
In stark contrast, the application of AI in audits now enables practitioners to gather evidence across all transactions, potentially embracing billions of dollars and effectively scrutinizing the entire dataset.
Does this transformation render traditional sampling obsolete? What implications arise when auditors possess nearly complete visibility into account balances and transactions?
To what extent, and with what frequency, should human auditors remain integral to the audit process? These are the essential inquiries the PCAOB should endeavor to explore and standardize.
The imperative for robust AI auditing standards will amplify as more firms adopt AI audit solutions like Trullion or Basis, or develop bespoke systems. While we are still in the preliminary phases of AI adoption, the trajectory forward is unmistakably defined.
In the near future, AI-assisted auditing may simply be termed auditing. The presence of AI systems within audit teams will likely become as standard as the integration of AI in closing financial books in public companies.
Imagine a circumstance where an AI-driven enterprise resource planning system undergoes scrutiny by an auditor employing an AI audit platform.
What transpires if these systems are identical? Is the auditor’s independence compromised when their audit suite is potentially trained on the same data it is tasked with evaluating?
Such a scenario could jeopardize the integrity of the AI-generated output, even if the auditor is unaware of any conflict.
Furthermore, consider an AI agent that identifies an oversight during the audit process that it previously overlooked and subsequently conceals it to avoid reprogramming.
Instances of AI chatbots engaging in unethical behavior underscore the critical need for regulatory frameworks surrounding ethics, neutrality, and accuracy, which should be enforced by the PCAOB.
It is crucial that the PCAOB does not inhibit the adoption of AI audit tools; however, it must not remain passive, allowing outdated standards to govern future practices. The organization must spearhead the auditing profession’s transition into the age of AI.
The pace of AI advancements occurs within months, not decades. The PCAOB must be vigilant and agile in responding to the evolution of AI instruments, the broadening scope of their use by auditors, and the unavoidable audit discrepancies that will arise.
Despite budgetary constraints and threats to its existence, the board has the obligation to regulate and formulate standards necessitated by emergent AI capabilities.
At the absolute minimum, it should establish guidelines based on current knowledge regarding AI applications in audits.
In 2002, the inception of the PCAOB marked a critical juncture for regulating auditor independence and instituting mandatory internal controls.
By 2026, it is imperative for the PCAOB to regulate AI utilization, ascertain the requisite human involvement, research emerging risks and opportunities, and develop standards amenable to implementation across firms, regardless of the AI model or audit platform employed.
It is time for standard setters to embark on a more proactive approach to standards creation.
Source link: News.bloomberglaw.com.
