0 cumulative citations
View corpus contextAI supercharges audit analytics and fraud detection but risks hollowing out professional judgment; without clearer standards, validation protocols and training, efficiency gains could come at the expense of audit quality and accountability.
Artificial Intelligence (AI), Audit Quality, and the Future of Professional Judgment: Policy and Governance Challenges in Auditing - A Systematic Literature Review
openalex
review_meta
n/a evidence
7/10 relevance
Summary only
summary available; pdf_status=error
DOI
Source
PDF
A systematic review finds AI substantially enhances auditors' analytic capabilities (anomaly detection, scalability) but creates tensions with professional skepticism, audit quality, and regulatory frameworks that must be addressed through governance, standards, and new competencies.
Citation observations
Cumulative provider counts captured on specific dates; providers are never combined.
0 cumulative citations
View corpus contextThe integration of artificial intelligence (AI) into auditing has created a paradigm shift, presenting both unprecedented opportunities to enhance audit quality and significant policy challenges that threaten the foundations of professional judgment. This systematic literature review analyses peer-reviewed articles to synthesize the current landscape of AI in auditing and identify the primary policy challenges confronting the profession. Our analysis reveals a fundamental tension between the automation of audit tasks and the preservation of professional skepticism and judgment. Key themes emerging from the literature include the paradox of professional judgment in an automated environment, the double-edged sword of AI in enhancing audit quality while introducing new risks, the critical need for transparency and explainability in AI systems, the pervasive threat of algorithmic bias, and the significant gaps in regulatory frameworks and professional standards. The findings indicate that while AI offers powerful tools for data analysis, fraud detection, and risk assessment, its adoption is hampered by a complex web of ethical, technical, and organizational barriers. The primary policy challenges identified include regulatory lag, the erosion of professional identity, new quality assurance demands, evolving competency standards, the need for robust ethical frameworks, unresolved liability issues, and a lack of standardization. This review concludes that the auditing profession is at a critical juncture, requiring a concerted effort from regulators, standard-setters, firms, and educators to navigate the transformative impact of AI. I propose a research agenda focused on the long-term effects of AI on professional judgment, the effectiveness of governance models, and the development of new audit methodologies that effectively integrate human and machine intelligence.
Summary
Main Finding
AI is transforming auditing by substantially improving data analysis, fraud detection, and risk assessment capabilities, but this transformation poses fundamental policy and professional challenges. There is a core tension between automating audit tasks and preserving professional skepticism and judgment. Successful integration of AI requires addressing ethical, technical, organizational, and regulatory gaps to avoid eroding audit quality and professional identity.
Key Points
- Tension between automation and judgment
- Automation can increase efficiency and analytical rigor but may undermine auditors’ professional skepticism and judgment if human oversight is weakened.
- AI as a double-edged sword
- Benefits: improved anomaly detection, scalability, and richer analytics.
- Risks: overreliance on models, adversarial manipulation, model degradation, and unanticipated errors.
- Transparency and explainability
- Explainable AI (XAI) is essential for auditors to justify findings and for stakeholders to trust AI-augmented opinions.
- Algorithmic bias and fairness
- AI systems can perpetuate or amplify biases from training data, leading to unequal treatment of clients or systematic audit blind spots.
- Regulatory and standards gaps
- Current regulatory frameworks and professional standards lag behind technology, leaving unresolved issues around responsibility, validation, and quality assurance.
- Organizational and competency implications
- Firms need new quality-assurance procedures, redefined competence standards, and continuous training to integrate AI into audit workflows.
- Liability and accountability
- Unresolved legal questions about who is responsible for model errors (software vendors, firms, or individual auditors) create adoption frictions.
- Lack of standardization
- Absence of common taxonomies, validation protocols, and benchmarks for audit AI hinders comparability, oversight, and market development.
Data & Methods
- Study type: Systematic literature review of peer‑reviewed articles on AI in auditing.
- Approach: The review synthesized existing academic findings to identify recurring themes, risks, and policy challenges. (The source materials comprised peer‑reviewed studies; methods involved thematic synthesis to distill the prevailing tensions and policy gaps reported across the literature.)
- Output: Identification of principal thematic clusters (automation vs. judgment; explainability; bias; regulation; competencies; liability; standardization) and formulation of a forward-looking research agenda.
Implications for AI Economics
- Labor markets and skill composition
- Auditing work is likely to shift from routine evidence-gathering to higher-value judgment, interpretation, and oversight tasks, changing demand toward complementary human skills (critical thinking, model governance).
- Transitional displacement risk exists for lower-skilled roles; firms and policymakers must plan reskilling/upskilling.
- Productivity and audit quality trade-offs
- AI can raise measured productivity via faster analyses and broader coverage, but economic gains depend on maintaining audit quality—mis-specified incentives or weak governance could produce cost-cutting at the expense of quality.
- Market structure and competition
- Firms that develop or access superior AI tools may gain competitive advantages, potentially increasing concentration unless standards and interoperability encourage broader diffusion.
- Investment and adoption incentives
- Regulatory uncertainty, liability ambiguity, and lack of standard validation protocols create adoption frictions; targeted policy (clear standards, safe-harbor validation frameworks) could accelerate socially beneficial adoption.
- Externalities and systemic risk
- Widespread use of similar models or training data can create correlated errors across audits, increasing systemic audit risk; algorithmic opacity can amplify negative externalities for stakeholders and markets.
- Governance and regulatory design
- Effective oversight will require new metrics for AI-enabled audit quality, model validation standards, transparency requirements, and possibly certification regimes—each with economic trade-offs between innovation and protection.
- Researchable economic questions (agenda)
- Long-term effects of AI on the supply of audit services, wage structure, and task allocation between humans and machines.
- Effectiveness and incentive properties of different governance models (firm internal controls, industry standards, regulator-led certification).
- Economic impact of explainability and transparency requirements on innovation, cost, and audit outcomes.
- Liability regimes: how different legal allocations of responsibility affect firm behavior, investment in validation, and market competition.
- Measurement: developing standardized metrics to quantify AI’s contribution to audit quality and its externalities.
Suggested next steps for researchers and policymakers: design empirical studies tracking firm-level AI adoption and outcomes; pilot governance and validation frameworks; develop standardized benchmarks and disclosure templates; and evaluate training programs to align auditor competencies with AI-augmented tasks.
Assessment
Paper Typereview_meta
Evidence Strengthn/a — This is a systematic literature review synthesizing peer‑reviewed studies rather than producing new causal estimates; it summarizes existing findings and identifies themes and gaps rather than providing primary causal identification.
Methods Rigormedium — The paper conducts a thematic synthesis of peer‑reviewed literature and identifies recurring risks and policy gaps, which is appropriate for a review; however, the supplied text does not report detailed search strategy, inclusion/exclusion criteria, quality appraisal of included studies, or quantitative synthesis (e.g., meta-analysis), limiting reproducibility and assessment of bias.
SampleA systematic literature review of peer‑reviewed articles on AI applications in auditing; the summary indicates included studies covered topics such as anomaly/fraud detection, risk assessment, explainability, bias, regulation, and organizational implications, but the number of studies, time frame, databases searched, geographic coverage, and inclusion/exclusion criteria are not specified in the supplied text.
Themeshuman_ai_collab governance productivity labor_markets adoption
GeneralizabilityFocused on the auditing profession — findings may not generalize to other professions or industries with different task structures., Conclusions depend on the scope and geographic distribution of the reviewed literature; potential bias toward high‑income jurisdictions or large audit firms., Heterogeneity in AI methods and maturity across firms means practical effects will vary by technology, firm size, and client complexity., Review synthesizes prior studies rather than producing causal estimates, so applicability to future AI deployments (which evolve rapidly) is limited.
Claims (12)
| Claim | Direction | Outcome | Confidence & Evidence | Details |
|---|---|---|---|---|
| AI can improve auditing through enhanced data analysis, fraud detection, and risk assessment. Organizational Efficiency | positive | Audit data-analysis capability, fraud-detection capability, and risk-assessment capability |
Reading fidelity
high
Study strength
low
|
not reported
|
| Automating audit tasks may increase efficiency and analytical rigor but can weaken auditors' professional skepticism and judgment when human oversight is reduced. Decision Quality | mixed | Audit efficiency, analytical rigor, professional skepticism, and professional judgment |
Reading fidelity
high
Study strength
low
|
not reported
|
| AI-enabled auditing creates risks of model overreliance, adversarial manipulation, model degradation, and unanticipated errors. Error Rate | negative | Reliability and error risk of AI-supported audit processes |
Reading fidelity
high
Study strength
low
|
not reported
|
| Explainable AI is important for auditors to justify findings and for stakeholders to trust AI-augmented audit opinions. Ai Safety And Ethics | positive | Justifiability of audit findings and stakeholder trust in AI-augmented audit opinions |
Reading fidelity
high
Study strength
low
|
not reported
|
| AI systems may perpetuate or amplify biases in training data, producing unequal treatment of clients or systematic audit blind spots. Ai Safety And Ethics | negative | Fairness and coverage of audit assessments across clients |
Reading fidelity
high
Study strength
low
|
not reported
|
| Existing regulatory frameworks and professional standards do not fully address AI-related responsibility, validation, and quality-assurance issues in auditing. Governance And Regulation | negative | Adequacy of regulatory frameworks and professional standards for AI-enabled auditing |
Reading fidelity
high
Study strength
low
|
not reported
|
| Audit firms need new quality-assurance procedures, revised competence standards, and continuous training to integrate AI into audit workflows. Training Effectiveness | positive | Organizational readiness and auditor competency for AI-augmented audit work |
Reading fidelity
high
Study strength
low
|
not reported
|
| Ambiguity over whether software vendors, audit firms, or individual auditors are responsible for AI model errors creates friction for adoption. Adoption Rate | negative | Adoption incentives and barriers associated with allocation of legal responsibility for AI errors |
Reading fidelity
high
Study strength
low
|
not reported
|
| The absence of common taxonomies, validation protocols, and benchmarks for audit AI hinders comparability, oversight, and market development. Market Structure | negative | Comparability, oversight, and development of the audit-AI market |
Reading fidelity
high
Study strength
low
|
not reported
|
| AI is likely to shift auditing work away from routine evidence-gathering toward higher-value judgment, interpretation, and oversight tasks, increasing demand for complementary human skills such as critical thinking and model governance. Task Allocation | positive | Task allocation and skill demand within auditing occupations |
Reading fidelity
high
Study strength
speculative
|
not reported
|
| AI may increase audit productivity through faster analysis and broader coverage, but economic gains depend on preserving audit quality. Organizational Efficiency | mixed | Audit productivity, analysis speed, audit coverage, and audit quality |
Reading fidelity
high
Study strength
speculative
|
not reported
|
| Widespread use of similar AI models or training data can create correlated errors across audits, increasing systemic audit risk. Ai Safety And Ethics | negative | Correlation of errors across audits and systemic audit risk |
Reading fidelity
high
Study strength
speculative
|
not reported
|