What role artificial intelligence can play (or not) in auditing
The continuous evolution of technology, hyper-automation, machine learning or deep learning and artificial intelligence are changing more and more fields of activity.
Now, with these technologies, enormous volumes of data can be analyzed to uncover anomalies and identify trends, patterns, and relationships that might otherwise escape human detection. Despite this, human understanding and experience remain crucial to understanding the results, determining the authenticity of the anomalies, and interpreting their implications in a larger context.
However, the collaboration between artificial intelligence and human auditors represents a solution with multiple advantages. First, AI can take over repetitive and laborious tasks, allowing human auditors to focus on the more complex and analytical aspects of the audit. This leads to increased efficiency and reduced time required to complete the audit process.
Accessibility to artificial intelligence is changing the auditor’s role
Although not all auditors and not all firms currently have access to the specialists to develop personalized AI-based tools, the many software solutions available have made its adoption more accessible. Auditors can choose the best option based on their requirements, resources and schedule.
As AI-based technologies advance, access will increase even further, enabling more auditors and firms to provide increased value to clients and encouraging standardization. Integrating AI into auditing streamlines data processing tasks, allowing auditors to focus on analysis and assessment.
This change reduces the time spent gathering, correlating, formatting and synthesizing information. Auditors can now drill down into data, gain valuable insights early in the audit process, and have a more efficient audit approach.
However, the change brought about by the accessibility of AI also requires an adaptation of auditors’ skills. They must understand emerging technologies, be able to interpret the results provided by algorithms, and navigate an ever-changing environment.
The challenges of using artificial intelligence in auditing
The use of artificial intelligence in auditing brings with it many challenges, although it brings significant benefits in the efficiency and accuracy of the process. One of the major challenges is the lack of transparency of algorithms, especially in the case of deep neural networks. These “black boxes” can complicate understanding how artificial intelligence reaches certain conclusions, an essential aspect in financial auditing where the ability to explain decisions is crucial.
Another challenge is the risk of bias, that is, biases or unwanted associations introduced into the data models used to train AI algorithms. The data used to train these algorithms may unconsciously reflect certain biases in society. If these biases are not identified and corrected, they can affect the objectivity of the audit process, generating incorrect or discriminatory results.
Cyber security is also a major challenge. Highly sensitive and confidential financial data is at increased risk of unauthorized access or manipulation when managed by AI algorithms. Implementing robust security measures becomes crucial to protecting information integrity.
Confidence in what artificial intelligence “works”
Advanced artificial intelligence tools such as neural networks are an undeniable force in analyzing and interpreting large volumes of data. The ability of these algorithms to learn complex patterns and extract meaningful information from large data sets brings significant benefits to various fields, from medicine to finance to scientific research.
However, as the use of these tools increases, concerns about the lack of transparency become more pronounced. Deep neural networks, in particular, are often perceived as “black boxes” where decision-making processes are difficult to understand and interpret. This raises important issues regarding ethics, accountability and trust in artificial intelligence systems.
Lack of transparency can make it difficult to explain how an algorithm reached a certain conclusion or decision. This aspect is crucial, especially in fields such as health or finance, where a clear understanding of the decision-making process is essential to justify and validate the results.
Solving this transparency issue is essential to promote user confidence it of artificial intelligence. Improving decision-making processes, as well as adopting ethical and accountability standards in the development and implementation of these technologies, can help manage concerns about the lack of transparency in the field of artificial intelligence. This would allow the full benefit of the capabilities of these technologies without compromising the integrity and trust in the decision-making process.
Therefore, AI can make biased or inaccurate predictions if it is trained on data with a high bias or inadequate data. Addressing this challenge requires management and auditors to implement appropriate controls, ensuring the accuracy and correctness of the AI’s “work” results.
Confidence in the accuracy of artificial intelligence remains a significant obstacle to its widespread adoption, underscoring the need for transparency and traceability in applications based on this technology.
Considerations for management and auditors
To ensure the effectiveness of AI-based audit tools, management and auditors should consider the following:
- Can management explain and evaluate the results of AI tools, guaranteeing the completeness, accuracy and effectiveness of internal controls?
- Can auditors explain and evaluate the results of AI audit tools, ensuring that they have obtained sufficient appropriate audit evidence to form an opinion?
- What are the basic requirements for understanding the native programming, controls, and processes around maintaining AI tools?
As artificial intelligence continues to transform the audit profession, auditors and firms must rise to the challenges, build trust and continually adapt to change. By embracing transparency, implementing effective controls and improving understanding of AI-based tools, auditors can harness the full potential of technology to deliver greater value in the digital age.
Alina Făniță este CEO și Partener al PKF Finconta. A lucrat cu companii multinaționale sau firme antreprenoriale din domenii diverse de activitate, pentru a le oferi servicii de audit financiar, due diligence, restructurări de grupuri, audit intern și alte servicii conexe activității de control intern. Este membră a celor mai prestigioase asociații profesionale din domeniu: ACCA (Association of Chartered Certified Accountants), CECCAR (Corpul Experților Contabili și Contabililior Autorizați din România), CAFR (Camera Auditorilor Financiari) și IIA (Institute of Internal Auditors). A absolvit EMBA Asebuss la Kennesaw State University, a fost trainer pentru cursuri IFRS și este invitată ca expert la numeroase conferințe de business. alina.fanita@pkffinconta.ro
