Daniel E. O'LEary
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Daniel E. O'LEary has a total of 349 co-authors. Here are the top co-authors by number of shared publications:
Veda C. Storey
Georgia State University
3 shared publicationsGuido Geerts
Lerner College of Business & Economics, Department of Accounting, University of Delaware, Newark,Delaware, USA
2 shared publications- Kevin C. Moffitt
Rutgers, The State University of New Jersey
2 shared publications Marcia Weidenmier Watson
Trinity University
2 shared publications- Yangin Yoon
Seoul National University of Science and Technology Seoul South Korea
2 shared publications - A. Faye Borthick
Georgia State University
1 shared publication
Latest Publications
- Open Access
- 7 October 2026
- Published by SAGE Publications in Journal of Information Technology Teaching Cases
The next generation of business intelligence: Large language models, model context protocol, and enterprise Bedrock
In late 2024, Anthropic released the “Model Context Protocol” (MCP) which allows users and developers to directly integrate frontier AI models with existing data and software tools. MCP is an open-standard protocol that is now being used with leading edge large language models to help with business intelligence. This case drills down on how AI is now integrated into BI software and on the use of the MCP in two settings: at the individual and at the enterprise level. Its use is illustrated on a sample problem. Although there are benefits with the integration of AI and BI, AI usage introduces other sources of costs, risks, and organizational effects that need to be examined. This case also provides the opportunity to investigate these and other issues in the context of BI software in a cloud computing setting. This case is based on field research and aimed at advanced undergraduates, graduate students, or executives in analysis of data and analytics or artificial intelligence issues.
- 7 October 2026
- Published by SAGE Publications in Journal of Information Technology Teaching Cases
Open AccessThe next generation of business intelligence: Large language models, model context protocol, and enterprise Bedrock
In late 2024, Anthropic released the “Model Context Protocol” (MCP) which allows users and developers to directly integrate frontier AI models with existing data and software tools. MCP is an open-standard protocol that is now being used with leading edge large language models to help with business intelligence. This case drills down on how AI is now integrated into BI software and on the use of the MCP in two settings: at the individual and at the enterprise level. Its use is illustrated on a sample problem. Although there are benefits with the integration of AI and BI, AI usage introduces other sources of costs, risks, and organizational effects that need to be examined. This case also provides the opportunity to investigate these and other issues in the context of BI software in a cloud computing setting. This case is based on field research and aimed at advanced undergraduates, graduate students, or executives in analysis of data and analytics or artificial intelligence issues.
- 1 July 2026
- Published by Wiley in Intelligent Systems in Accounting, Finance and Management
- Vol. 33 (3)
Some Large Language Model Applications at the Big Four: Demonstration‐Based Field Work
Recent headlines indicate that the Big Four accounting, auditing, and professional services firms (Big Four) have invested billions in artificial intelligence (AI), as they digitally transform their firms. As part of that analysis, this paper reviews several large language model (LLM) applications that have been developed and considers some of the implications of those applications, such as the democratization of knowledge and the potential impact on barriers to entry. Because LLMs are early in their life cycle and there is limited information about those applications, I take a demonstration‐based field work approach, gathering information from firm demonstrations and then building some prototypes in LLMs that simulate the demonstrated capabilities.
- 1 July 2026
- Published by Wiley in Intelligent Systems in Accounting, Finance and Management
- Vol. 33 (3)
Some Large Language Model Applications at the Big Four: Demonstration‐Based Field Work
Recent headlines indicate that the Big Four accounting, auditing, and professional services firms (Big Four) have invested billions in artificial intelligence (AI), as they digitally transform their firms. As part of that analysis, this paper reviews several large language model (LLM) applications that have been developed and considers some of the implications of those applications, such as the democratization of knowledge and the potential impact on barriers to entry. Because LLMs are early in their life cycle and there is limited information about those applications, I take a demonstration‐based field work approach, gathering information from firm demonstrations and then building some prototypes in LLMs that simulate the demonstrated capabilities.
- 27 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Intelligent Systems
- Vol. 41 (3) , 110-114
Large Language Model Commerce
So-called “Chatbots,” powered by large language models (LLMs) initially were part of the effort aimed at developing systems that could provide human-like conversations. However, LLMs continue to add capabilities to perform various additional activities. This paper introduces the notion of “Large Language Model Commerce” and discusses the evolving nature of LLM capabilities. Search and knowledge management capabilities are turning large language models into potential e-commerce machines. Google Trends and Wikipedia page views are used empirically to analyze emerging implications of that evolution.
- 27 April 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Intelligent Systems
- Vol. 41 (3) , 110-114
Large Language Model Commerce
So-called “Chatbots,” powered by large language models (LLMs) initially were part of the effort aimed at developing systems that could provide human-like conversations. However, LLMs continue to add capabilities to perform various additional activities. This paper introduces the notion of “Large Language Model Commerce” and discusses the evolving nature of LLM capabilities. Search and knowledge management capabilities are turning large language models into potential e-commerce machines. Google Trends and Wikipedia page views are used empirically to analyze emerging implications of that evolution.
- 6 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Intelligent Systems
- Vol. 41 (1) , 113-117
Large Language Model Use of Text Information: Trouble With Semantic Underspecification
This article investigates the ability of large language models (LLMs) to successfully answer questions about text when there is “semantic underspecification.” Although humans make substantial use of semantic underspecification, and seem to be able to tease out meaning, LLMs do not successfully respond to underspecified text with the same level of correctness. We found whereas humans were able to disambiguate underspecified text, LLMs required additional information specification to correctly answer queries. These findings have several implications.
- 6 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Intelligent Systems
- Vol. 41 (1) , 113-117
Large Language Model Use of Text Information: Trouble With Semantic Underspecification
This article investigates the ability of large language models (LLMs) to successfully answer questions about text when there is “semantic underspecification.” Although humans make substantial use of semantic underspecification, and seem to be able to tease out meaning, LLMs do not successfully respond to underspecified text with the same level of correctness. We found whereas humans were able to disambiguate underspecified text, LLMs required additional information specification to correctly answer queries. These findings have several implications.
- 2 January 2026
- Published by Taylor & Francis in Journal of Decision Systems
- Vol. 35 (1)
Do users anchor on large language model rankings? How do they adjust?
- 2 January 2026
- Published by Taylor & Francis in Journal of Decision Systems
- Vol. 35 (1)
Do users anchor on large language model rankings? How do they adjust?
- 8 September 2025
- Published by American Accounting Association in Journal of Emerging Technologies in Accounting
- Vol. 22 (2) , 125-147
Exploratory Experiments in Generating Bag of Word Dictionaries Using Word2Vec and ChatGPT
This paper investigates the use of machine learning as a means of generating “bag of words” using text corpora from accounting applications. We use Word2Vec to generate words/dictionaries, that are “similar” to a seed word that captures a concept. As part of our analysis, we perform several experiments using text from Form 10Ks and earnings calls. We investigate several activities including choice of the seed word(s), choosing word sources (corpora), analysis of resulting word lists, and other concerns. We also examine the notion of “human-in-the-loop” and the roles that a person would need to perform while generating a dictionary. Further, we investigate the impact of using accounting and financial corpuses on the different semantic and syntactic relationships, in contrast to Wikipedia. We then extend the analysis to compare those findings to ChatGPT another source of words and investigate some of the advantages and disadvantages of that approach. Data Availability: Data are available from the authors. JEL Classifications: M4.
- 8 September 2025
- Published by American Accounting Association in Journal of Emerging Technologies in Accounting
- Vol. 22 (2) , 125-147
Exploratory Experiments in Generating Bag of Word Dictionaries Using Word2Vec and ChatGPT
This paper investigates the use of machine learning as a means of generating “bag of words” using text corpora from accounting applications. We use Word2Vec to generate words/dictionaries, that are “similar” to a seed word that captures a concept. As part of our analysis, we perform several experiments using text from Form 10Ks and earnings calls. We investigate several activities including choice of the seed word(s), choosing word sources (corpora), analysis of resulting word lists, and other concerns. We also examine the notion of “human-in-the-loop” and the roles that a person would need to perform while generating a dictionary. Further, we investigate the impact of using accounting and financial corpuses on the different semantic and syntactic relationships, in contrast to Wikipedia. We then extend the analysis to compare those findings to ChatGPT another source of words and investigate some of the advantages and disadvantages of that approach. Data Availability: Data are available from the authors. JEL Classifications: M4.
- 14 July 2025
- Published by Wiley in Intelligent Systems in Accounting, Finance and Management
- Vol. 32 (3)
Editorial: Analysis of Sentiment Estimates and Cognitive Fallacies in Large Language Models
This paper describes some experimentation with the evolving ability of large language models to generate sentiment estimates. We find that current models seem to equal or even exceed the ability of human annotators in a case study of single sentiment sentences. In addition, using the large language models, we were able to identify a small number of sentences in the data set, where it appears that the annotator made errors in assessing the sentiment. Unfortunately, analysis of the LLM results also illustrates apparent cognitive biases in the LLM behavior. Those effects appear to include an “ostrich effect” and a “no one is good enough” effect cognitive bias in LLM sentiment estimates.
- 14 July 2025
- Published by Wiley in Intelligent Systems in Accounting, Finance and Management
- Vol. 32 (3)
Editorial: Analysis of Sentiment Estimates and Cognitive Fallacies in Large Language Models
This paper describes some experimentation with the evolving ability of large language models to generate sentiment estimates. We find that current models seem to equal or even exceed the ability of human annotators in a case study of single sentiment sentences. In addition, using the large language models, we were able to identify a small number of sentences in the data set, where it appears that the annotator made errors in assessing the sentiment. Unfortunately, analysis of the LLM results also illustrates apparent cognitive biases in the LLM behavior. Those effects appear to include an “ostrich effect” and a “no one is good enough” effect cognitive bias in LLM sentiment estimates.
- Open Access
- 15 June 2025
- Published by Association for Computing Machinery (ACM) in Digital Government: Research and Practice
- Vol. 6 (2) , 1-13
Using Large Language Models for Armchair Auditors
Armchair auditors are citizens who use open data to investigate and monitor government activities, typically using analytics and other approaches. Armchair auditors provide a valuable role in holding governments and organizations accountable. This paper investigates the potential use of large language models (LLM) to support armchair auditor analyses of different governmental entities. Unfortunately, the literature, prior to the development of LLM suggested several challenges for armchair auditors. However, the analysis in this paper suggests that LLM can provide substantial data and analytic process support for armchair auditors mitigating issues such as providing guidelines for analyses, guiding users to appropriate communities, suggesting potential data availability opportunities, doing analysis, and other issues. As part of an approach to unifying armchair auditor searches, this paper also suggests a prompt library designed to support, standardize and promote best practice analyzes among armchair auditors. In addition to these issues, this paper also analyzes emerging ethical issues associated with armchair auditors and their use of open data and LLMs. Finally, this paper extends the activity theory model to account for LLMs.
- 15 June 2025
- Published by Association for Computing Machinery (ACM) in Digital Government: Research and Practice
- Vol. 6 (2) , 1-13
Open AccessUsing Large Language Models for Armchair Auditors
Armchair auditors are citizens who use open data to investigate and monitor government activities, typically using analytics and other approaches. Armchair auditors provide a valuable role in holding governments and organizations accountable. This paper investigates the potential use of large language models (LLM) to support armchair auditor analyses of different governmental entities. Unfortunately, the literature, prior to the development of LLM suggested several challenges for armchair auditors. However, the analysis in this paper suggests that LLM can provide substantial data and analytic process support for armchair auditors mitigating issues such as providing guidelines for analyses, guiding users to appropriate communities, suggesting potential data availability opportunities, doing analysis, and other issues. As part of an approach to unifying armchair auditor searches, this paper also suggests a prompt library designed to support, standardize and promote best practice analyzes among armchair auditors. In addition to these issues, this paper also analyzes emerging ethical issues associated with armchair auditors and their use of open data and LLMs. Finally, this paper extends the activity theory model to account for LLMs.
- 10 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Intelligent Systems
- Vol. 40 (2) , 23-26
An Anchoring Effect in Large Language Models
This article examines a well-known bias in human reasoning. Anchoring occurs when people receive a piece of information that influences their future judgments. Human use of anchoring biases has been well studied; however, the purpose of this article is to examine the existence of anchoring within large language models (LLMs). Experiments were performed on three different LLMs, and each LLM demonstrated an anchoring bias when asked to provide numerical estimates about text information. It was determined that LLMs will directly attend to information provided to them and, like humans, anchor their judgments on that information. As the manipulated experimental judgment is only one of several potential dimensions, the extent to which the anchoring bias propagates to the other decision dimensions was also examined.
- 10 April 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Intelligent Systems
- Vol. 40 (2) , 23-26
An Anchoring Effect in Large Language Models
This article examines a well-known bias in human reasoning. Anchoring occurs when people receive a piece of information that influences their future judgments. Human use of anchoring biases has been well studied; however, the purpose of this article is to examine the existence of anchoring within large language models (LLMs). Experiments were performed on three different LLMs, and each LLM demonstrated an anchoring bias when asked to provide numerical estimates about text information. It was determined that LLMs will directly attend to information provided to them and, like humans, anchor their judgments on that information. As the manipulated experimental judgment is only one of several potential dimensions, the extent to which the anchoring bias propagates to the other decision dimensions was also examined.
- 7 April 2025
- Published by American Accounting Association in Current Issues in Auditing
- Vol. 19 (1) , A1-A9
Data-Driven Audits: Audit Analytic Platforms and General Ledger Analytic Tools
SUMMARY: The future of auditing is a data-driven audit, which extracts and analyzes digital data stored in client accounting systems. To help conduct data-driven audits, audit firms are developing audit analytic platforms (AAPs). This paper focuses on one key application tool of AAPs that analyzes the entire population of general ledger transactions. These general ledger analytics tools (GLATs) use assessment routines to assign a risk score to each transaction, helping guide the auditor’s planning, risk assessment, and selection of high-risk transactions for substantive procedures. GLATs help support audit quality and provide valuable new insights about client operations. We discuss the analytics routines used by GLATs, the forces driving their use, and some potential limitations of this technology. Data Availability: Data are available from the authors. JEL Classifications: M4.
- 7 April 2025
- Published by American Accounting Association in Current Issues in Auditing
- Vol. 19 (1) , A1-A9
Data-Driven Audits: Audit Analytic Platforms and General Ledger Analytic Tools
SUMMARY: The future of auditing is a data-driven audit, which extracts and analyzes digital data stored in client accounting systems. To help conduct data-driven audits, audit firms are developing audit analytic platforms (AAPs). This paper focuses on one key application tool of AAPs that analyzes the entire population of general ledger transactions. These general ledger analytics tools (GLATs) use assessment routines to assign a risk score to each transaction, helping guide the auditor’s planning, risk assessment, and selection of high-risk transactions for substantive procedures. GLATs help support audit quality and provide valuable new insights about client operations. We discuss the analytics routines used by GLATs, the forces driving their use, and some potential limitations of this technology. Data Availability: Data are available from the authors. JEL Classifications: M4.