Historically, literature suggests that a variety of accounting roles will be replaced by Artificial Intelligence (AI) and related technologies; however, in recent years there is a growing recognition that accounting can in fact harness AI’s potential to add value to organisations. Commentators have highlighted the need for increased research exploring accounting and AI and for accounting scholars to consider multi-disciplinary research in this area. This study uses a form of topic modelling to analyse literature exploring AI and related techniques in an accounting context. Latent Dirichlet Allocation (LDA) has been used to enable probabilistic, machine-based interrogation of large volumes of literature. This study applies LDA to the abstracts of 930 peer-reviewed academic publications from a variety of disciplines to identify the most significant accounting and AI topics discussed in the literature during the period 1990 to 2023. Our findings suggest that prior literature reviews based on more traditional methodologies do not capture a comprehensive picture of accounting and AI research. Eleven topic clusters are identified which provide a comprehensive topology of the extant literature discussing accounting and AI and set out an agenda for future research designed to foster academic progress in the area. It also represents one of the first applications of probabilistic topic modelling to accounting literature.
In March 2022, the Securities and Exchange Commission (SEC) proposed the mandatory reporting of cybersecurity risk management policies for public companies. This study aims to explore the potential impact of cybersecurity risk management strategy disclosure on nonprofessional investors. Using a 4 x 1 between-participants experimental design, we examine whether nonprofessional investors’ perceptions and decisions differ between disclosed cybersecurity risk management strategies of self-assessment, self-assessment referencing the National Institute of Standards and Technology (NIST) framework, third-party assurance, and insurance. We find that nonprofessional investors’ willingness to invest is significantly higher for the insurance strategy compared to the third-party cybersecurity examination and self-assessment (without reference to NIST) strategies. Moderated mediation analysis reveals that investors’ perceptions of financial risk moderates the mediating effect of perceived cybersecurity risk management strategy effectiveness on the relation between cybersecurity risk management strategy and likelihood of investment. Our study contributes to regulators, practitioners, and stakeholders concerned about the potential impact of cybersecurity risk management strategy disclosures on nonprofessional investors.
This paper proposes the concept of artificial intelligence co-piloted auditing, emphasizing the collaborative potential of auditors and foundation models in the auditing domain. The paper discusses the future relationship and interactions of human auditors and AI, imagining an audit setup where auditors’ capabilities are enhanced through artificial intelligence across a variety of audit tasks. To exemplify the potential of this co-piloted audit paradigm, we illustrate a systematic fine-tuning approach to foundation models using Chain-of-Thought prompting. This study showcases how foundation models can work as collaborators flexibly with auditors, enabling the model to accurately identify transactions from instructions. This study provides a detailed description of the formulated prompt protocols and the corresponding responses generated by ChatGPT, ensuring reproducibility. We envision this work as an initial step towards the widespread implementation of co-piloted auditing, paving the way for more efficient, accurate, and insightful audit procedures.
To address the need for reporting and disclosure of cryptocurrency holdings in compliance with the FASB guidance for the use of fair value measurements for cryptocurrency (FASB, 2023), this paper develops a modeling process for reporting entities to measure the market value of cryptocurrencies with limited or no observable transactions. In this valuation model, we consider the last observable market information with time decay, its comparable assets market index, and dynamic real-time market participants’ sentiment and attention. Notably, the application of exogenous variables allows us to maximize the observable inputs in measuring fair value, such as asset classification based on economic traits and market participants’ attention and sentiment measurement with online media textual analytics. We propose a valuation framework and construct a prediction model that can achieve a prediction accuracy of 87 % on target asset resurging prices.