Abstracts

Sustainability and Environmental, Social, and Governance (ESG) reporting have assumed growing importance in capital markets, particularly in shaping investor decision-making. However, the continuing absence of universally accepted reporting standards and domain-specific assurance grants managers substantial discretion to present their sustainability performance in a deceptively favorable way, which is commonly referred to as greenwashing. This risk is further compounded by investors’ limited access to robust data and specialized analytical expertise, which impairs their ability to detect greenwashing in ESG disclosures. Large Language Models (LLMs), with advanced reasoning capabilities and integrated web search functions, offer a novel approach to addressing these challenges. This study proposes a framework for developing LLM-based agentic systems to detect greenwashing risks. Since greenwashing is not a binary outcome but a multi-dimensional phenomenon, this study introduces seven greenwashing indicators to assess different facets of greenwashing risks. Those indicators are further integrated into a dashboard tailored to the needs of various users. This study contributes to the greenwashing literature by establishing a comprehensive framework for constructing LLM-based agents to detect greenwashing and proposing a set of quantifiable greenwashing indicators. It also contributes to the accounting profession by providing a tool for real-time monitoring of greenwashing risks.

Keywords: Large Language Model (LLM); greenwashing; sustainability; Environmental, Social, and Governance (ESG); Artificial Intelligence (AI); agentic system

Corporate carbon disclosure, including Scope 3 emissions, has significantly increased in recent years and is increasingly important for regulators, investors, and stakeholders. Scope 3 data remains challenging to quantify due to reliance on fragmented, secondary sources, leading to significant information gaps and unreliable disclosures. This study introduces an innovative mobile application framework to enhance Scope 3 emissions data collection and management. Our app-based AI-driven estimation addresses these shortcomings through interactive digital tools and gamification to capture granular and accurate data. By gathering primary data directly from consumers, employees, and suppliers, our app significantly enhances data quality, accuracy, and credibility, fostering richer, more transparent sustainability disclosures. This improved data availability empowers firms to meet reporting obligations and provides stakeholders with trustworthy information, potentially enhancing ESG ratings and market reputation. Beyond improved disclosures, the mobile app facilitates real-time sustainability management through continuous feedback. Management can quickly identify high-emission areas and make informed decisions, fostering operational efficiencies, cost savings, and innovation. The framework also addresses practical challenges in stakeholder engagement, data reliability, verification, and privacy. This paper contributes to sustainable accounting and accounting information systems literature by presenting a scalable, modular solution for enhanced Scope 3 emissions measurement, improved ESG transparency, and effective response to evolving regulatory and market demands. The mobile app facilitates real-time, stakeholder-driven data collection, significantly enhancing corporate decision-making and sustainability management.

Keywords: ESG, Scope 3 Emissions, Mobile Application, Sustainability Accounting, Carbon Disclosure

This study presents a new approach to the content analysis of ESG reports, with a primary focus on manual content analysis based on the inductive identification of seven ESG themes. The analysis was conducted through detailed manual coding, and the themes were constructed based on the content of the reports without predefined categories. This approach allowed for the creation of a clear structure of ESG themes aligned with the EU's Corporate Sustainability Reporting Directive (CSRD) and the European Sustainability Reporting Standards (ESRS) guidelines. Artificial intelligence was used solely to test scalability with a larger dataset, which included 289 reports from publicly listed companies in the EU. The use of AI ensured that the manually derived themes could be identified and categorized within a large and diverse dataset. This study demonstrates that although ESG report analysis is a multistep process, manual content analysis provides a strong foundation upon which effective and scalable analytical tools can be integrated for in depth content examination and reliable result generation.

Keywords: Content analysis, Artificial intelligence, Text mining, Theme extraction

Prior research has documented a number of positive outcomes attributable to managerial ability, such as higher earnings quality, decreased likelihood of going concern opinions, and shorter audit report lags. In this study, we examine the association between managerial ability and cybersecurity breaches, a growing concern for a wide range of stakeholders. We posit that higher managerial ability translates to an improved set of cybersecurity risk mitigation measures, thereby lowering the likelihood of cybersecurity breaches. Using the Demerjian et al. (2012) measure of managerial ability on a sample of 13,813 firms, results indicate that higher managerial ability is more likely to attenuate cybersecurity breach incidences. Considering the effects of managerial ability on internal and external breaches separately, we further find that higher managerial ability is effective at mitigating risks of both breach types. This study contributes to the literature by providing empirical evidence of an association between managerial ability and cybersecurity breaches, for both internal and external breach types.

Keywords: Managerial Ability, Cybersecurity, Internal Breach, External Breach

This study conducts a comparative analysis of the information content embedded in mandatory cybersecurity disclosures and their implications for firm financial performance. Focusing on U.S. public companies from 2019 to 2021, the analysis introduces a novel metric—Disclosure Completeness Intensity (DCI)—to evaluate the depth and completeness of cybersecurity-related disclosures within two mandated sections of annual reports: Risk Factors (RF) and Management’s Discussion and Analysis (MDA). The findings reveal a negative association between disclosure completeness and future financial performance, suggesting that more comprehensive disclosures may signal heightened underlying risks. Furthermore, cybersecurity disclosures within the RF section exhibit greater informational value than those in the MDA section, as evidenced by a higher incidence of statistically and economically significant relationships with forward-looking performance indicators. These effects are amplified when the sample is stratified by industry based on historical cybersecurity breach frequency. This study provides the first empirical evidence differentiating the informativeness of topic-specific cybersecurity disclosures between RF and MDA. The findings offer important regulatory insights, particularly in the context of the SEC’s 2023 final rule on cybersecurity risk disclosure, and contribute practical recommendations for enhancing disclosure guidance. Finally, the study highlights fruitful avenues for future research in topic-specific disclosure and cybersecurity governance.

Keywords: Financial Performance, Cybersecurity risk disclosure; Risk Factors, Management’s Discussion and Analysis, SEC 2023

This study examines the SEC’s Item 1C cybersecurity disclosure, focusing on its disclosure quality and whether it enhances corporate cybersecurity governance. Building on prior voluntary disclosure practices under Item 1A, the 2023 mandate introduces structured reporting requirements concerning firms' cybersecurity risk management, board oversight, and managerial responsibilities in Item 1C. Using advanced textual analysis, I measure disclosure quality by specificity, applying document embeddings and cosine similarity metrics to categorize disclosures. The analysis shows that Item 1C disclosures are more standardized and generalized compared to Item 1A disclosures. To assess the disciplinary effect of the mandate, I implement a difference-in-differences (DID) design that leverages variation in compliance timing across public and SRC filers. The results show that mandatory cybersecurity disclosures are associated with a significant reduction in breach incidence only for firms that demonstrate a deeper understanding of sector-specific cybersecurity threats and governance expectations, as reflected in their industry-specific risk factor disclosures. Additional robustness tests support the validity of these findings. Overall, the evidence suggests that the SEC’s Item 1C mandate contributes to improved cybersecurity outcomes when firms provide disclosures that meaningfully reflect their industry-specific risk environment.

Keywords: Cybersecurity; SEC disclosure mandate; breach incidents; disclosure disciplinary effect; textual analysis; disclosure quality.

In the context of pressure to ensure real-time data transparency and reporting, the accounting profession must increasingly make judgments about using unfamiliar sources. In contrast to prior studies that typically focus on data analytics rather than how big data is judged and used, this study investigates how accounting professionals use big data for business-related decision-making. In exploring, we conduct 24 semi-structured interviews with accounting professionals, using a vignette for a consistent context as they make six decisions requiring use of accounting information and big data as information sources. Through application of valence theory to quantify the data, findings show how their propositional judgments positively value accounting information and negatively value most characteristics of big data. Moreover, when invoking a practical judgment about whether and how to use big data for decision-making, these professionals apply a heuristic, and weigh its ‘relevant-value-to-context’ (RVC) by anchoring their judgment in accounting information. In correlating these findings with their expressed trust and preference to rely on experience when decision-making, practical judgment appears to be informed by experience. As both judgments value accounting information, participants expressed trust in big data appears to relate to their willingness to accept risk rather than assurance about its truth or ability. Accordingly, findings suggest a theoretical need to revisit the role of experience and trust in the extended valence framework. Further, whereas practitioner resources focus on how big data’s management value chain affects accountants’ roles and required skills, findings show the importance of existing capabilities, particularly their training and experience.

Keywords: Accounting information: big data; valence theory; decision-making; judgment

While traditional machine-readable datasets (such as Compustat) have advanced financial capital markets research, a significant amount of valuable data remains locked in unstructured, PDF-based financial disclosures. This limitation hampers the scalability of archival research. Prior approaches relying on keyword-based heuristics have struggled to handle the variability and complexity of financial reporting formats. This study proposes an AI-based scalable framework for extracting structured data from unstructured financial reports. Leveraging a context-aware prompting strategy with large language models (LLMs), the framework identifies relevant page ranges, isolates targeted disclosure sections, and extracts structured quantitative information. To demonstrate its practical utility, we apply the framework to retrieve capitalization thresholds disclosed in the “Capital Assets” section of the note disclosures within governmental Annual Comprehensive Financial Reports (ACFRs). An evaluation on a sample of county and municipal ACFRs yields 99% accuracy in identifying relevant page ranges and 95% accuracy in extracting data at the individual note level. The study contributes to (1) the accounting literature by enabling the construction of novel datasets for large-sample archival research, (2) the AIS literature by demonstrating how emerging technologies can address complex data extraction challenges, and (3) practice by enhancing the efficiency of regulatory oversight and public sector monitoring.

Keywords: Large Language Models, Financial Disclosures, Unstructured Data, Information Extraction, Natural Language Processing, Government financial reports