IASB Chair: AI Must Not Hollow Out Professional Judgment in Financial Reporting
Outgoing IASB Chair Andreas Barckow used his final conference address to put a pointed question to the global accounting profession: what happens to professional judgment when AI handles the work through which judgment is built? Delivered at the IFRS Foundation Conference on 29 June 2026, the speech is not a warning against technology. It is a call for deliberate thinking about how firms train, supervise, and develop people in an environment where automation is expanding faster than governance frameworks can keep up.
The Context: A Tenure Defined by Disruption
Barckow's term at the IASB ended on 30 June 2026. Over that period, the themes he returned to most often were uncertainty, complexity, and fragmentation. Geopolitical pressures reshaped the conditions in which companies operate and report. Those conditions, he argued, are no longer temporary. They have become structural features of the environment.
Against that backdrop, he chose his final address to turn attention to a different kind of disruption: the rapid spread of generative and agentic AI across every sector, including financial reporting.
What AI Can Do, and What It Cannot
Where AI Adds Genuine Value
Barckow acknowledged that the IASB has already seen practical benefits from AI in its own processes. Disclosure analysis is one clear example. AI can scan large volumes of financial reports against specific disclosure requirements and assess how those requirements are being applied in practice globally. Work that previously required significant time now completes in a fraction of it.
Speed has value. But in financial reporting, speed alone is not the standard. Information must be robust, reliable, and trusted. That distinction matters, and it sits at the heart of his concern.
The Limits: Nuance, Exceptions, and Inconsistency by Design
Accounting might appear well-suited to AI. It operates within principles, concepts, and rules. But Barckow was direct about the complexity that sits beneath that surface. IFRS Standards have been developed over fifty years by different people, in different economic environments, responding to different stakeholder pressures. Some requirements exist specifically because they were needed to achieve consensus or get a standard across the line. There are exceptions, exemptions, policy choices, and deliberate inconsistencies between standards that serve a purpose.
That is not a flaw in the system. It reflects the reality that standard-setting is a human process, shaped by context and compromise. AI trained on those standards inherits that complexity without necessarily understanding the reasoning behind it. The risk of misapplication, or of accepting an AI output that looks plausible but is subtly wrong, is real.
For firms working on crypto compliance reporting, where the interaction between evolving digital asset guidance and existing IFRS or US GAAP frameworks already demands careful interpretation, that risk is amplified. Crypto financial statements sit at the intersection of several standards, and the judgment calls involved are not mechanical.
The Deeper Problem: How Judgment Is Built
Experience as the Raw Material of Expertise
Barckow's sharpest point was about professional development. Judgment, he argued, is not instinctive. It is built through repeated exposure: performing analysis, making mistakes, correcting them, challenging information rather than accepting it, and building the pattern recognition that eventually tells an experienced professional when something deserves a second look.
Much of that experience is accumulated through the more mechanical, ground-level work that junior professionals have traditionally done. If AI takes over that work, the efficiency gain is visible and immediate. What is lost is less visible but potentially more consequential: the process through which the next generation learns to exercise judgment.
A Question Firms Need to Answer Now
Barckow did not claim to have solutions. He was explicit about that. But he framed the questions clearly: How do firms train, supervise, and develop people when AI is handling the tasks through which judgment has historically been formed? How do accounting firms and audit practices ensure that junior professionals still encounter the difficult, ambiguous situations that build expertise?
These are not abstract questions for standard-setters alone. They are operational questions for every accounting firm, CFO function, and audit practice integrating AI into its workflows right now. The same applies to firms handling blockchain analytics data quality reviews, where the combination of novel asset types and limited precedent means that human oversight of AI outputs is not optional.
Three Cornerstones That Have Not Changed
Reflecting on his time leading the IASB, Barckow identified three principles that have become more important as conditions have become more volatile, not less.
Listening: The One Task AI Cannot Assume
For a global standard-setter, listening is not a courtesy. It is a functional requirement. The IASB serves a diverse ecosystem spanning large and small jurisdictions, preparers and investors, auditors and regulators. Stakeholders want to explain their positions in their own language and from their own context. When people feel genuinely heard, and when the reasoning behind a decision is explained clearly even if the outcome is not what they wanted, trust in the institution grows.
Barckow was clear that this is the part of standard-setting that automation cannot replace. Human interaction and reflection are not inefficiencies in the process. They are the process.
Communication: Focused, Not Exhaustive
Standard-setters, he observed, tend to assume that stakeholders follow every project and understand every decision. That assumption is wrong. Stakeholders are managing their own information overload. The answer is not silence. It is focus: communicating what matters, why it matters, and how it affects the businesses and professionals in the audience.
For firms monitoring developments in areas like professional education standards or evolving crypto IFRS accounting guidance, the volume of regulatory output is already significant. Clear, targeted communication from standard-setters reduces the risk of firms missing material developments buried in broad consultation documents.
Focus and Prioritisation: Less Is Still More
Barckow returned to a point he made four years earlier at the same conference: less is more. A focused work plan allows the IASB to give each project the time and attention it requires. It also gives stakeholders a realistic opportunity to engage. He said he believes that principle more strongly now than when he first articulated it.
The implication for standard-setting is that the number of projects completed is not a measure of quality. The depth of engagement with each project is. That has direct relevance for anyone watching the IASB's agenda on digital asset accounting, where the interaction between IFRS crypto assets guidance and broader fair value measurement standards continues to evolve.
What This Means for Accounting Firms and CFO Functions
Barckow's speech was addressed to the standard-setting community, but the operational implications extend well beyond it.
Firms integrating AI into their accounting and audit workflows face a version of the same challenge the IASB faces: how to use technology to increase efficiency without degrading the human judgment that quality depends on. In practice, that means several things.
Supervision structures need to account for the fact that junior staff may be reviewing AI outputs rather than building analysis from scratch. That is a different cognitive task, and it may require different training and oversight models. Firms cannot assume that exposure to AI-generated work provides the same developmental foundation as doing the underlying work manually.
For crypto financial statements specifically, the complexity is compounded. The standards governing digital asset accounting, whether under IFRS or US GAAP frameworks like ASC 350-60, require judgment calls that sit outside the mainstream of what AI has been trained to handle reliably. Fair value measurement for volatile assets, classification decisions, and disclosure adequacy all require contextual reasoning that experienced professionals develop over time.
The firms that will navigate this well are those that treat AI as a tool requiring human oversight, not a system that can operate independently, and that invest in the professional development structures that preserve judgment capacity across generations of staff.
FAQs
What did outgoing IASB Chair Andreas Barckow say about AI and accounting?
In his final address as IASB Chair, Barckow acknowledged that AI offers real efficiency gains for tasks like disclosure analysis, but warned that over-automating the work through which junior professionals build experience could erode the professional judgment that high-quality financial reporting depends on. He asked whether the profession has thought carefully enough about that risk.
Does this speech change any IFRS standards or requirements?
No. This was a reflective keynote address, not a standard-setting announcement. It does not alter existing IFRS requirements, including those relevant to crypto assets or digital asset accounting. The speech signals priorities and concerns rather than regulatory change.
What are the practical implications for accounting firms using AI tools?
Firms need to review how AI integration affects the development of junior staff. If AI handles the analytical work through which professionals traditionally built judgment, firms need alternative mechanisms to ensure that experience and expertise continue to develop. Supervision structures, training programmes, and review processes all require re-examination.
How does this relate to crypto financial statements and digital asset accounting?
Crypto financial statements and related disclosures under IFRS or US GAAP involve judgment calls in areas like classification, fair value measurement, and disclosure adequacy. These are precisely the areas where AI outputs require careful human review, and where the judgment built through experience is most critical. Firms should not assume AI can handle these determinations reliably without oversight.
What are the three cornerstones Barckow identified for high-quality standard-setting?
Listening carefully to a diverse range of stakeholders, including smaller jurisdictions and those whose views might otherwise be overlooked; communicating decisions clearly and explaining the reasoning behind them; and maintaining a focused, prioritised work plan rather than pursuing volume for its own sake.
Source: IFRS Foundation