![\"Business](\"https:\/\/www.europesays.com\/ai\/wp-content\/uploads\/2026\/06\/1781520190_355_0x0.jpg\")
<\/p>\n<\/p>\n
Business professional concentrates in coworking space leveraging AI powered virtual assistant and realtime big data analytics on laptop display for hyper automated workflow intelligence optimization<\/p>\n
getty<\/p>\n
AI is already reshaping audit risk faster than accounting standards can respond. Firms are deploying systems that classify transactions, draft workpapers and surface exceptions, yet no authoritative standard defines what adequate human oversight looks like or how to allocate responsibility when AI gets the numbers wrong.<\/p>\n
Imagine an accounting AI agent that classifies 10,000 transactions overnight, drafts the workpapers and flags three anomalies for review. A staff accountant checks the flagged items and signs off. Six months later, a restatement reveals that the AI missed a pattern of misclassified revenue \u2014 the kind an experienced accountant would have caught but that never appeared in the system\u2019s exception report. The result is an ethics gap the profession\u2019s current standards don\u2019t fully address.<\/p>\n
So who is responsible? The staff accountant who reviewed only what the system surfaced. The firm that deployed a tool it believed was reliable. Or the vendor whose terms of service disclaim accuracy altogether. After years working in the spaces where these rules get made, including a term on the U.S. Federal Government Cost Accounting Standards Board, I can say this with confidence: Under today\u2019s ethics codes, the answer isn\u2019t clear \u2014 and that uncertainty should concern anyone whose business decisions rest on AI\u2011initiated audited financial statements.<\/p>\n
Conduct Codes Built For A Different World<\/p>\n
The accounting profession has one of the most developed ethics architectures of any field. The AICPA Code of Professional Conduct<\/a>, the international code maintained by the IESBA and oversight bodies like the PCAOB set clear expectations around competence, due care and professional skepticism<\/a>.<\/p>\n Every one of those frameworks assumes a human is doing the work. “Due professional care” envisions a person exercising judgment. Documentation requirements assume someone can explain the reasoning behind a conclusion. AI doesn’t work that way. It produces outputs, not always full explanations. When those outputs are wrong, the existing rules don’t clearly allocate responsibility.<\/p>\n The IESBA revised its ethics code, effective December 2024, to address technology but used the generic term “technology” throughout, with nothing specific to AI\u2019s opacity or tendency to hallucinate. The AICPA has published AI guidelines for forensic and valuation work but labeled them explicitly as “neither authoritative guidance nor standards.” As practitioners at a 2024 AICPA webcast on AI in audits<\/a> observed, there is still no authoritative AICPA standard governing AI use in practice.<\/p>\n ForbesThe Accounting Skill AI Cannot ReplaceBy Yvonne Hinson<\/a>The Crumple Zone<\/p>\n Researcher Madeleine Clare Elish uses the term \u201cmoral crumple zone\u201d<\/a> to describe what happens next. In a car, the crumple zone absorbs impact to protect the vehicle’s structure. In human-AI systems, the human absorbs blame to protect the integrity of the technology, even when they had limited control over its behavior.<\/p>\n The accounting profession is building a possible crumple zone out of its junior staff. They review AI output and carry professional liability when something breaks, even when they may lack the experience to know what “wrong” looks like in that context. That liability just became more personal: The PCAOB lowered its contributory liability standard<\/a> from recklessness to negligence, meaning individuals face consequences for contributing to violations even without intent.<\/p>\n The errors are real. By late 2025, researchers had documented nearly 800 cases of AI-generated hallucinations across 25 countries \u2014 fabricated citations, invented regulations and confident references to sources that don\u2019t exist. In tax practice, AI models hallucinate form references, cite outdated law and produce answers that look correct but aren\u2019t. These errors look authoritative on the page. Under time pressure, a reviewer may never catch them.<\/p>\n AI vendors disclaim accuracy in their terms of service. Major insurers are introducing AI-specific exclusions<\/a> to errors and omissions policies, which means the professionals bearing the liability may not even have coverage for it from the vendor.<\/p>\n Where Regulation Is Headed<\/p>\n The SEC charged two investment advisers in 2024<\/a> for misrepresenting their AI capabilities, the first “AI-washing” enforcement actions. Its Investor Advisory Committee has since recommended<\/a> that companies disclose how they deploy AI and how boards oversee it. These are early signals, but show where this is headed.<\/p>\n Firms are investing in internal AI governance, and many take it seriously. However, internal frameworks don\u2019t define liability, mandate oversight or carry enforcement consequences. The profession needs authoritative standards, and it needs them before the first major AI-assisted audit failure forces the question in a courtroom rather than a standard-setting body.<\/p>\n Three Questions The Profession Must Answer<\/p>\n The academic world training new accountants is moving quickly to ensure AI is understood and handled properly. AACSB\u2019s framework for AI in business education identifies “responsible AI and ethics integration” as critical, and 97% of accredited schools now address AI. <\/p>\n At the American Accounting Association, we\u2019re pushing to ensure ethics keeps pace with adoption, not trail behind it. However, we have seen our own issues with papers submitted to journals that contain hallucinated references. <\/p>\n The profession won’t close this gap until it answers three questions its current codes don’t address: <\/p>\n What counts as adequate human oversight of AI-assisted work? When AI contributes to a material error, how should responsibility be allocated? What obligation does a firm have to evaluate an AI tool’s reliability before deploying it on an engagement?<\/p>\n These are not only academic questions. They are the foundation on which the financial system operates. The profession that addresses them first may establish the standards for other professions and industries confronting similar challenges. The profession that hesitates may find its future defined by decisions made elsewhere.<\/p>\n Yvonne Hinson, PhD, CPA is CEO of the American Accounting Association and a former executive with the Association of International Certified Professional Accountants. A longtime Wake Forest University professor, department chair, and Dean of the Charlotte campus, she has held key academic and professional board roles and was named one of the Most Powerful and Influential People in Accounting (2024, 2025).<\/p>\n","protected":false},"excerpt":{"rendered":"Business professional concentrates in coworking space leveraging AI powered virtual assistant and realtime big data analytics on laptop…\n","protected":false},"author":2,"featured_media":74139,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[2],"tags":[40875,24,40876,1798,25,40877,23947,1528,40874,1795,40878],"class_list":["post-74138","post","type-post","status-publish","format-standard","has-post-thumbnail","category-ai","tag-accounting-ai","tag-ai","tag-ai-accounting","tag-ai-governance","tag-artificial-intelligence","tag-audit-risk","tag-audits","tag-ethics","tag-iesba","tag-responsible-ai","tag-when-ai-makes-mistakes"],"_links":{"self":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/74138","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/comments?post=74138"}],"version-history":[{"count":0,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/posts\/74138\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media\/74139"}],"wp:attachment":[{"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/media?parent=74138"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/categories?post=74138"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.europesays.com\/ai\/wp-json\/wp\/v2\/tags?post=74138"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}