跳到正文
原文
The CPA Journal(美国 CPA 技术期刊)· Khondkar E. Karim, DBA, CPA·· 2026-10-05精选站内评分56

内部审计师与外部审计师采用人工智能的认知与体验差异研究

The Varied Perceptions and Experiences of Internal Versus External Auditors in Adopting Artificial Intelligence

导读

一项访谈3名内部审计师和5名外部审计师的研究显示,两类审计师均有意采用人工智能,但内部审计师的实际使用经验更丰富。外部审计师认为人工智能可加快任务和测试并降低成本,同时担忧入门级岗位替代、数据可靠性及与客户互动减少。受访者认为组织需提供人才和培训支持,并准备验证人工智能输出及审计人工智能技术本身。

推荐理由

比较8名内外部审计人员的认知与采用意愿,呈现AI落地中效率收益、输出验证和人员培训等共同挑战。

正文

Audit firms have invested heavily in artificial intelligence (AI) in recent years, with media outlets reporting billions of dollars in collective investments from the Big Four alone (Big4Events, “How Deloitte, KPMG, EY, and PwC Are Leading the AI Transformation,” November 24, 2025, https://www.big4events.com/ai-revolution-in-big-4-consulting). For example, KPMG employs IBM Watson’s AI-powered systems to analyze the creditworthiness of commercial mortgage loan portfolios. Deloitte uses AI solutions from Kira Systems to review documents such as contracts, invoices, leases, and tweets. A growing stream of academic literature has studied the potential impact of different types of AI technologies on the current and potential future state of the audit profession. For example, survey evidence indicates significant efficiency gains from the adoption of robotic process automation in Big Four firms (L. A. Cooper, D. K. Holderness, T. L. Sorensen, and D. A. Wood, “Robotic Process Automation in Public Accounting,” Accounting Horizons, vol. 33, no. 4, pp.15-35, 2019). Similarly, the use of drones and automated counting software improves both audit and financial reporting quality (M. H. Christ, S. A. Emett, S. L. Summers, and D. A. Wood, “Prepare for Takeoff: Improving Audit Efficiency and Effectiveness with Drone-Enabled Inventory Audit Procedures,” Review of Accounting Studies, vol. 26, pp.1323-1343, 2021).

To better understand how AI is used in audit practices, this article investigates CPAs’ existing knowledge about AI as a general technology, their intentions to adopt AI in auditing, and the factors that may have contributed to such intentions. Moreover, it explores potential differences in responses between internal and external auditors, given that they face significantly different types of audit objectives and tasks. The case study research method was used to interview eight auditing professionals (Exhibit 1), three internal auditors, and five external auditors in order to gain insights on these issues. Case studies use multiple sources of evidence (e.g., interviews with different individuals) to investigate a contemporary phenomenon and are particularly useful when an existing body of knowledge is insufficient to permit the posing of causal questions (A. J. Mills, G. Durepos, and E. Wiebe, Encyclopedia of Case Study Research, Sage Publications, 2009; R. K. Yin, Case Study Research: Design and Methods, Sage Publications, 2003). The authors chose the case study approach because the purpose of this investigation is exploratory rather than confirmatory.

EXHIBIT 1

Interviewee Demographics

 Interviewee; Highest Degree; Firm Size; Audit Firm Expertise; External or Internal; Position 1; Master’s; 81; All (assurance, taxation, consulting, and other); External; Partner or Equivalent 2; MBA; 4,000; N/A; Internal; Chief Audit Officer 3; Bachelor; 775; All; External; Senior Manager 4; Bachelor; >220; All; External; Partner or Equivalent 5; MSA; >10,000; All; External; Manager 6; MBA, MSA; >400; N/A; Internal; Department Leader 7; MBA; 1; Consulting; External; Director 8; Bachelor; >10,000; N/A; Internal; Senior Officer

The audit professionals that were interviewed are employed by various types of institutions—both for-profit and nonprofit firms of varying sizes in different industries. The auditors included range from the manager level to the partner (or equivalent) level. Exhibit 1 shows the demographics of the interviewees.

Thematic analysis was then employed in order to identify recurring themes that emerged from interviews, which is outlined below (V. Braun and V. Clarke, “Using Thematic Analysis in Psychology,” Qualitative Research in Psychology, vol. 3, pp.77–101, 2006). As part of this process, it is important to highlight the differences that were observed in the perceptions and responses of internal versus external auditors.

Auditors’ Understanding of AI

The interviewees had different levels of knowledge about AI, ranging from merely being aware of it to fully understanding the uses and potential of these technologies. Specifically, one external auditor had limited knowledge about AI, “very rudimentary at best,” and at the time of the interview, their firm had just begun “looking at data analytics and how to implement it into our audit process.” Another external auditor “did not go too in-depth into it.” The remaining interviewees (including both internal and external auditors) had varying levels of somewhat deeper understanding of AI and its potential in auditing.

Several interviewees (both internal and external auditors) indicated that their firms used some type of AI program, and sometimes outsourced such needs, if it was an advanced program. External auditors who worked (or have worked) at smaller accounting firms stated that they did not use such programs because their firms did not need sophisticated AI systems.

Perception of Using AI in Audit Practices

Interviewees were asked how AI may affect audit risk and audit procedures. Responses indicated that AI influences how firms deploy human auditors, raising the question of how to define the role of human auditors moving forward. In addition, external auditors said that AI helps to streamline audit engagement. Both internal and external auditors suggested that AI can be simply viewed as a tool, but they also recognized uncertainties about what type of tool it can be (e.g., whether it should be a management tool or an auditing tool). For example, one external auditor noted that, “it’s really important that we look at the use of technologies as an input into our processes and not the driver and certainly not the determinant of what we do.”

When asked how AI can change the effectiveness or efficiency of audit engagements, external auditors said that AI can help complete tasks faster or engage in tasks (such as testing) that may no longer need to be delegated to human auditors. Although they perceive that using AI may cause issues related to software glitches, they agreed that AI is more reliable and makes less mistakes compared to human auditors. Both internal and external auditors expressed concerns about exactly how effectiveness and efficiency gains will be achieved. Most of the conversations were at a higher level, as implementing and streamlining AI into the audit process had yet to be accomplished (e.g., “How do we get auditors to become comfortable with such tools?”). Specifically, an external auditor expressed a concern that the use of AI “is going to be hamstrung by the regulatory environment … you might be able to convince your local audit partner on your engagement to rely on your usage of AI but are you going to convince Deloitte?” Compared to the external auditors, the internal auditors seemed to have more experience using AI in their work, indicating that they have adopted AI quite extensively. For example, one internal auditor indicated that their firm has embedded AI in the internal controls of the firm, and one such application is detecting employees’ potential misconduct.

In addition to greater efficiency and effectiveness, both internal and external auditors indicated that AI could improve quality and reduce costs. AI technologies also may have spillover effects, allowing auditors to be more strategic and to focus on issues using the information generated by AI. This idea resonated among the external auditors, presumably given that these professionals have traditionally served in functions or roles heavily involving manual, time-consuming tasks. The internal auditors interviewed also expressed a similar view, although they did not believe that AI adoption would drastically change the nature of their current responsibilities, since their work tended to involve the use of emerging technologies already.

External auditors cited three potential unintended consequences of adopting AI. First, AI may replace the use of entry-level auditors, changing the audit workforce. Second, data generated by AI may lack reliability, and auditors may not know how to assess its reliability. Third, AI may reduce human interaction, so the auditors will not get to know their clients.

Finally, responses were mixed when interviewees were asked how comfortable they or their peers were with AI. Most (if not all) of the interviewees expressed being comfortable with AI, even if they may not know at an operational level how to program or use it. Some said generally employees are receptive to at least learning more about AI and possess the ability and skills to do so. External auditors primarily stated that many are still not comfortable with AI (e.g., because of a skills gap) and thus, training is vital to acclimate employees to AI.

Intentions to Adopt AI

Most of the interviewees expressed the explicit intention to use AI now or in the future. Both internal and external auditors stated that the AICPA has provided a lot of information about AI for auditors, convincing many that it will be a disruptive technology. Both groups cited benefits of adopting AI, including aggregating information and detecting issues such as fraud, and indicated that new technology is inevitable and natural, and that they intend to adopt AI. One external auditor suggested that the size of the firm matters: larger firms have more resources and may have greater intentions to adopt AI.

Some interviewees expressed concerns related to using AI and about how humans will use the information continuously generated by AI. Similarly, other information-related challenges discussed include how to assess or verify AI-generated information, how detailed the information should be, the assumptions behind the information generated, and the costs associated with interpreting the data. For example, one internal auditor mentioned “Who’s looking at the information? Who’s accessing it? It’s your common challenges that you see with anything that’s extracting information on a periodic basis.” Both internal and external auditors expressed concerns about potential risks of using AI technology in auditing, such as how to ensure data integrity.

Organizational Support for Adopting AI

Responses about the future use of AI fall into two main categories. Internal auditors and external auditors from large firms felt that the “future is now,” as their firms are already using AI extensively. External auditors from smaller firms indicated that their firms will eventually adopt AI, but that they were uncertain about the timeline. This uncertainty seemed to be driven by firm needs as well as cost-benefit considerations. For example, no one expressed doubt that AI can help the firm, but some said that it may be costly to implement and streamline in terms of time, effort, and money.

Across both internal and external auditors, AI training does not seem to be a priority. But interviewees indicated that their employers do encourage and support professional development related to analytical skills such as data analytics and programming. For firms that do not use AI, training was even less of a concern. Overall, at the time of the interviews, firms were not requiring or compelling their auditors to become data scientists or AI specialists.

Firms may need to prepare for the unintended consequences of adopting AI, such as how to audit the AI technology itself and ensure the validity of the AI outputs.

Both internal and external auditors said that skills associated with using AI would be important when hiring staff in the future, though no one thought that it was necessary to overhaul current hiring practices, and said they focused on other areas when recruiting (e.g., information systems, computer science). Interviewees also indicated that regardless of the background of new recruits, employees will all receive training in the skills they need on the job, so new recruits need to have the ability to learn and work with others.

Outlook

Understanding auditors’ intentions to adopt AI and the underlying factors that drive such intentions is crucial for organizations and policymakers seeking to harness the potential of AI in auditing while addressing any associated challenges. These interviews provide insights into the perceived benefits and costs of AI for audit professionals as well as concerns about using AI in the audit.

For public accounting firms and internal auditors intending to use AI, this survey helps raise awareness of the necessary organizational support for adopting AI. Such support includes hiring the right personnel, providing additional training, and more importantly, enhancing the corporate readiness to adopt such technology. Moreover, firms may need to prepare for the unintended consequences of adopting AI, such as how to audit the AI technology itself and ensure the validity of the AI outputs.

Khondkar Karim, PhD, CPA, is a professor of accounting at the University of Massachusetts Lowell.

Karen Jingrong Lin, PhD, is a professor of accounting at the University of Massachusetts Lowell.

Cody Lu, PhD, is an assistant professor of accounting at the University of Massachusetts Lowell.

Hongwei Zhu, PhD, is the Rist Professor of Artificial Intelligence and a professor of information systems at the University of Massachusetts Lowell.

The post The Varied Perceptions and Experiences of Internal Versus External Auditors in Adopting Artificial Intelligence appeared first on The CPA Journal.

来源:The CPA Journal(美国 CPA 技术期刊) · cpajournal.com

审计热点 是聚合摘要和阅读索引,原文版权归各来源所有,引用请以官方原文为准。如果你是来源方,希望更正、下架或调整展示方式,可以通过反馈页联系我们。

来自微信公众号的内容,本站仅呈现摘要与原文链接,不转载文章全文;文章权利归属原作者及原公众号

关于使用条款隐私信源