ssrn-5511498

Artificial Intelligence, Human Intelligence and Corporate Governance Brian Bolton Professor of Finance Dwight W. Andrus, Jr. / BORSF Eminent Scholar Endowed Chair in Finance Moody College of Business University of Louisiana at Lafayette PO Box 43709 Lafayette, LA 70504 brian.bolton@louisiana.edu Jung Park Associate Professor of Entrepreneurship and Innovation Institut Supérieur de Gestion (ISG) Paris Campus Paris Ouest 8 rue de Lota, 75116 Paris jungeung.park@gmail.com June 2025 Keywords: artificial intelligence, machine learning, corporate governance, boards of directors, venture governance, family firms, corporate culture -- 1 of 28 -- ABSTRACT: The integration of Artificial Intelligence (AI) into corporate governance has begun to redefine how organizations operate and make strategic decisions. This study explores this integration of AI into corporate governance using a unique approach: more than 30 corporate governance experts were surveyed to obtain their diverse, expert perspectives on this issue. These corporate governance experts approached the future of AI and corporate governance from four distinct perspectives: cultural dimensions, venture governance, family firm governance and social impact governance. Despite working on vastly different research agendas, several common themes emerged from these contributors. Boards and executives are increasingly using AI tools for analytics, predictive modeling, and risk management, creating opportunities to enhance decision-making and transparency. However, this technological shift brings new challenges, including algorithmic bias, data privacy concerns, and ethical accountability. Corporate governance has always been a uniquely human function, whether that entails making decisions in the board room, creating governance systems and policies or debating shareholder proposals. Will AI allow these humans to serve their governance functions more efficiently and effectively or will it interfere with the rational, context-based decision-making that only humans provide? Artificial intelligence has become ubiquitous in most of our lives in recent years. But it is not new. The term ‘artificial intelligence’ was introduced in the 1950s, referring to the science of creating intelligent machines (Smulowitz and Vogel, 2024). Handheld calculators, fingerprint recognition and desktop computers were the early versions of applied AI, while machine learning and neural networks are the current, dynamic versions. AI has proceeded along a continuum of innovation for 70 years to get to where it is today. Corporate governance research has utilized elements of this innovation during the past few years to better understand how firms and leaders work to provide a return on investment to their stakeholders. Perhaps the most common way we have used advanced computing technology in our research is with textual analysis. Tim Loughran and Bill McDonald1 are responsible for much of this work, especially in finance and accounting research, creating dictionaries and programming algorithms to identify patterns and idiosyncrasies in corporate reporting (e.g. 10-K reports, press releases, earnings announcements). Actual language is compared to predicted or baseline language, to tease out sentiment that might be missed in traditional, discrete data collection. And this becomes the data sample. To date, that is largely how we have used this computing intelligence – to collect data. And the benefits have been enormous. We can collect large amounts of data very quickly, including data that we humans might not be able to objectively measure on our own.2 Once the code is written, and once the comparative dictionary is established, collecting the data becomes routine. The researcher still needs to analyze the data and reach conclusions or policy implications, but that may be changing soon, too. 1 See, for example, Tim Loughran and Bill McDonald, 2011, When is a Liability not a Liability? Textual Analysis, Dictionaries, and 10-Ks, Journal of Finance, 66:1, 35-65. 2 One of the authors studied CEO turnover as part of his dissertation in the early 2000s. He identified more than 2000 cases of CEO turnover and manually reviewed annual reports and press releases for each observation in order to categorize each turnover as some type of voluntary or disciplinary turnover. The entire process took nearly a month of full-time work, the type of work that only PhD students will do; today, a mere 20 years later, that same dataset probably could be collected in a day or two using relatively simple textual analysis. -- 2 of 28 -- Perhaps ironically, the study of AI does not rely on data; it currently relies on observing what is being done and how it is being use. After all, if corporate governance relates to the structures firms use and the decisions leaders make to provide a return on investment to suppliers of capital, then observing how AI relates to these structures and decisions seems to be the best place to start. For the corporate governance experts surveyed in this study, working to better understand how corporate governance structures are utilizing and impacted by AI- driven technology, this has created – and will continue to create – significant opportunities to advance corporate governance research. Having common themes – transparency, decision- making, organizational culture, regulation, ethical considerations – across different contexts allows us to formalize theories and normative perspectives. From these foundations, we can then apply, test and explain the confluence of corporate governance and AI using case studies, qualitative storytelling and even rigorous quantitative analysis. The opportunities for corporations to advance their governance systems using AI-driven technologies are enormous – including transparency, increased efficiency, more effective decision-making, and better stakeholder engagement. But, there are real concerns; the growing centrality of AI risks subordinating stakeholder trust to algorithms, overshadowing human strategic vision as well as ethical, moral, and social justice considerations. The ‘expert human touch’ will always be – and, we believe, must always be – the driving force behind any corporate governance system. Relative to many other streams of research, corporate governance research has uniquely focused on this human element. Corporate governance is about people. Our research attempts to understand their behaviors, incentives and decisions. And understanding these humans has always required sophisticated human intelligence from researchers. That will always be the case. But, the AI revolution has created a fascinating convergence of human intelligence and artificial intelligence – both for us as researchers and within the firms that we study. A few years ago, most of us could not imagine the impact that AI would be making on our lives today – and we could not imagine the impact that AI would be making within corporate boardrooms. Our job as researchers is to push our research into new directions and uncover new perspectives on corporate governance knowledge; and AI-driven technology has ignited enormous new opportunities for advancing our understanding of why businesses do what they do. The perspectives that these corporate governance experts advance in this study – at this intersection of human intelligence and artificial intelligence – will help shape what businesses do in the near future. References: Loughran, T,. and McDonald, B. (2011) When is a Liability not a Liability? Textual Analysis, Dictionaries, and 10-Ks, Journal of Finance, 66(1), pp. 35-65. Smulowitz, M., and Vogel, P. (2024). The Spandows: Pioneering AI in family philanthropy and sustainable business. In The Routledge Handbook of Artificial Intelligence and Philanthropy, pp. 275-286. Routledge. -- 3 of 28 -- Artificial Intelligence, Human Intelligence and Corporate Governance 1. Introduction In studying contemporary issues in corporate governance research, there are many novel dynamics that we need to consider: culture, international issues, start-up and venture governance, family firm governance, governance of transitioning firms, governance of mission and purpose and others. We see these issues as both unique and important issues for corporate governance research – and practice – both today and in the future. However, they are not the only issues worthy of deep exploration. Perhaps the most unique and important issue affecting corporate governance scholars, practitioners and regulators today is artificial intelligence (AI), just as it may be the most significant issue currently affecting all business. Thus, it is appropriate – and necessary – to build on these new directions and new perspectives on corporate governance research by exploring how the omnipresence of AI will impact corporate governance and what it means for scholars, practitioners and regulators in the future. In this exploration, we are taking a unique approach. We recently finished editing a book on new directions and perspectives on corporate governance research. Rather than choosing to provide a general survey of the research on corporate governance and artificial intelligence, or to address a singular novel question in depth, we decided to survey the authors to the book to get their perspectives on why AI matters to corporate governance and corporate governance research. More than 30 early-career academics, seasoned faculty, and practitioners have contributed to this book. These experts come from five continents and work in nearly 20 different countries. Some have already done their own novel research on AI and some have used it extensively in practice and research. It would be difficult to get more relevant, insightful, diverse and expert perspective on what the current state of AI in corporate governance is and how it will impact corporate governance going forward. This paper aims to provide the thoughts of these experts – plus our own contributions – from a variety of viewpoints. Our work as corporate governance researchers is largely an empirical field; yes, we have some foundational theories, as mentioned in the Introduction, but most of our work is advanced by studying how corporate governance systems are structured and the decisions that corporate leaders make. We are studying human behavior. In this paper, we want to think about how AI impacts this human behavior – and how it will impact the broader world of corporate governance in the future. As such, much of the discussion in this paper is about the practice of – not the research of – corporate governance. With this approach, our goal is to do what we promised we would do in the Introduction: trigger your imagination and lead you to ask important questions. By thinking about how corporate governance is practiced and how AI is being used in that practice, we can think of the questions we need to be asking in our own research. This paper will be largely conversational and anecdotal. Perhaps ironically, the study of AI does not rely on data; it currently relies on observing what is being done and how it is being use. After all, if corporate governance relates to the structures firms use and the decisions -- 4 of 28 -- leaders make to provide a return on investment to suppliers of capital, then observing how AI relates to these structures and decisions seems to be the best place to start. We begin this paper with a brief overview of how AI – and its related antecedents – has been used in corporate governance research to date. Then we explore several of the most significant themes related to using AI in corporate governance: strategic planning, decision-making, transparency, regulation and ethics. From there, we consider how AI relates to each of our four themes: cultural dynamics of corporate governance, start-up and venture governance, family firm governance and governance of mission and purpose. Again, most of the thought leadership in the book has been provided by our 30+ contributors; in some places, we use their thoughts to build an argument and in others we quote them directly. We conclude the paper by considering some opportunities for future research related to AI and corporate governance. We hope you enjoy this conversation as much as we enjoyed putting it together. 2. A brief history of artificial intelligence, machine learning and corporate governance research Artificial intelligence, or AI, has become ubiquitous in most of our lives in recent years. But it is not new. The term ‘artificial intelligence’ was introduced in the 1950s, referring to the science of creating intelligent machines (Smulowitz and Vogel, 2024). Handheld calculators, fingerprint recognition and desktop computers were the early versions of applied AI, while machine learning and neural networks are the current, dynamic versions. AI has proceeded along a continuum of innovation for 70 years to get to where it is today. Corporate governance research has utilized elements of this innovation during the past few years to better understand how firms and leaders work to provide a return on investment to their stakeholders. Perhaps the most common way we have used advanced computing technology in our research is with textual analysis. Tim Loughran and Bill McDonald3 are responsible for much of this work, especially in finance and accounting research, creating dictionaries and programming algorithms to identify patterns and idiosyncrasies in corporate reporting (e.g. 10-K reports, press releases, earnings announcements). Actual language is compared to predicted or baseline language, to tease out subtle cues or sentiment that might be missed in traditional, discrete data collection. And this becomes the data sample. To date, that is largely how we have used this computing intelligence – to collect data. And the benefits have been enormous. We can collect large amounts of data very quickly, including data – such as sentiment – that we humans might not be able to objectively measure on our own.4 Once the code is written, and once the comparative dictionary is established, 3 See, for example, Tim Loughran and Bill McDonald, 2011, When is a Liability not a Liability? Textual Analysis, Dictionaries, and 10-Ks, Journal of Finance, 66:1, 35-65. 4 One of the editors studied CEO turnover as part of his dissertation in the early 2000s. He identified more than 2000 cases of CEO turnover and manually reviewed annual reports and press releases for each observation in order to categorize each turnover as some type of voluntary or disciplinary turnover. The entire process took nearly a month of full-time work, the type of work that only PhD students will do; today, a mere 20 years later, that same dataset probably could be collected in a day or two using relatively simple textual analysis. -- 5 of 28 -- collecting the data becomes routine. The researcher still needs to analyze the data and reach conclusions or policy implications, but perhaps that will be changing soon, too. The biggest concern with this work is the measurement: in general, the data collection become less precise than it would be through other methods. Researchers have to be extremely thoughtful (and thorough) in establishing the data collection rules; they have to clearly specify both what they are looking for and in what context, as the programming will only collect what it is told to collect. For example, if we code to collect when a CEO refers to something as “best,” we need to be clear to exclude Best Buy (Loughran and McDonald, 2016). This requires significant upfront work on the part of the researcher, and also requires the researcher to critically evaluate what they have collected after the fact. We cannot allow these algorithms to become black boxes; we are responsible for the data we collect and the processes we use to collect them. Thus, we have continued to need to apply human intelligence to the evolving intelligence of the technology we use. And maybe this will always be the case. Or maybe we are now beginning to see the potential for artificial intelligence to replace human intelligence (or lack, thereof) in many research situations. ‘Machine learning’ is the term used to refer to advanced, dynamic artificial intelligence; the machine can dynamically update both the rules used to collect data and to modify outputs. If the machine knows that we are looking for something akin to ‘best financial performance,’ it will learn not to collect ‘Best Buy’ as an observation (or even something like ‘best manager’). As machines learn what we are trying to do, they can dynamically update what they are doing to better serve our purpose. In terms of efficiency, robustness and storytelling, the potential applications of AI and machine learning for most social science researchers have become extraordinary in recent years.5 Is this true for those of us who perform corporate governance research? In some ways, such as with finite, quantitative data-collection, it will certainly be true. But, ultimately, much corporate governance research is focused on how human beings behave. We are studying the decisions people make, the processes they used to make these decisions, how they interact with other human beings and what motivates them to do what is best. Early AI has relied on static rules and instructions, whereas all corporate governance revolves around unique and dynamic interactions and incentives. Certainly some of this can be teased out through the approaches discussed above. But can it all? Can AI tease out cultural differences, say, between China and the United States, as they impact a multinational firm’s corporate governance structure? Can AI help us clearly understand the dynamics between a sister and brother management team in a family business? Can AI help a start-up transition from garage to unicorn – and can it help us understand the unique incentives of founders and funders along this progression? Can AI dynamically interpret a non-profit’s mission to deliver meaningful implications about the organization’s behavior and purpose? Relative to many other streams of research, corporate governance research has uniquely focused on the human element. Corporate governance is about people; and people are unique. 5 To avoid bouncing back and forth between similar but different terms, for the balance of this paper we use the generic term “AI” to refer to this entire continuum of innovation, including artificial intelligence, textual analysis, machine learning, neural networks and anything else along this continuum. -- 6 of 28 -- We have tried to understand their behaviors, incentives and decisions. And understanding these humans has always required sophisticated human intelligence from researchers. Yet, twenty years ago, most of us could not imagine the impact that AI would be making on our lives today. Maybe the same will hold true for corporate governance research. We do not expect to come to many conclusions about the future of AI in corporate governance research in the sections that follow. We merely hope to identify new perspectives and new directions for corporate governance research to pursue. This exploration will begin by considering the practical application of AI in corporate governance activities, looking at how firms are – our could be – using AI in the governance. The purpose is to help us as scholars better understand the questions we should be asking – and then we conclude by directly addressing some of those questions and considering the research paths to pursue. We begin this survey and exploration thinking about the benefits and opportunities: AI enables rapid and critical analysis of vast amounts of data, enhancing strategic decision-making and implementation. It delivers innovative solutions, streamlines processes, boosts efficiency and productivity, and frees human resources for more value-added activities. (Cinzia Dessi) 3. Opportunities for strategic planning & decision-making From a practical perspective, AI is currently most common is strategic planning, decision- making and the day-to-day practice of corporate governance. As with any tool, before utilizing AI, corporate leaders need to know why they are using it and what they are trying to achieve with it: Whether an organization is for‐profit or non‐profit, embarking on an effective AI journey should start with identifying a specific need or challenge. Before delving into advanced AI techniques, it is important to grasp the basics. The quality of input data plays a pivotal role in achieving accurate AI results. Start with manageable goals to gauge the potential impact of AI. (Małgorzata Smulowitz) Once the goals and strategies are identified, how firms, leaders and boards of directors utilize AI to do their jobs has revolutionary potential. AI will transform corporate governance and its research by enabling corporations and boards to make more informed, data-driven decisions, more quickly. (Jing Zhao) Artificial Intelligence is set to transform corporate governance by enhancing decision- making, improving risk management, automating compliance monitoring, and fostering better stakeholder communication. By analyzing vast datasets, AI can provide predictive insights, reduce biases, and enable more informed and timely governance decisions. Its ability to proactively identify risks, streamline compliance processes, and deliver real-time updates to stakeholders will enhance transparency and operational efficiency. (Arjya Majumdar) -- 7 of 28 -- If we consider that transparency and accountability are the foundational priorities of any corporate governance system, we can imagine how AI will help organizations practice and implement corporate governance, beyond just utilizing it as a tool for enhancing operational efficiency and executing strategic planning. Another direct impact is how AI might change the way corporations govern their business, using AI as part of their governance mechanisms. This could be using generative AI tools for reporting, recruiting board directors or tools to monitor financial development. (Agnes Dyvik Clark) And, we can all imagine a situation where AI technology becomes a part of a firm’s governance structure – literally. That is, could AI become a participating, contributing and voting member of a board of directors? The answer appears to be ‘yes’ or ‘kind of.’ In early 2024, Abu Dhabi-based International Holding Company (IHC) appointed a non-voting board observer, ‘Aiden Insight’ (Aiden).6 Aiden’s role is to provide data analysis, risk assessment, compliance monitoring and strategic insights. In theory, that is exactly how every other company could be using AI-technology to support its board and leadership. What makes this appointment unique is that it is a board-level appointment. To IHC, this “is a clear indication of the company's forward-looking approach and its ongoing commitment to leveraging technology for sustainable growth and ethical excellence."7 Aiden does not need to be a voting member of the IHC board; and Aiden does not require compensation or stock options. We can all imagine the upside potential of such an appointment ; but how worried should we be about transparency, accountability and risk management associated with the decisions that Aiden makes ? We will return to these issues and potential concerns related to relying on AI in corporate governance later in this paper. We conclude this section with another case study example of how AI can be used in practice, authored by Léa Wang and Séverine Mulliez8 and from their perspective on brave communication in family firm governance. This case is based on a real-life example; “we imagined how this tragedy could have been avoided—or its impact reduced—if AI had been appropriately incorporated.” Perhaps utilizing AI in this manner, and as in the example above with Aiden at UHC, the nature of AI integration can dynamically improve itself, thus minimizing any inherent risks and improving upon natural human-to-human relationships. Now, imagine this: AI goes beyond mere note-taking. It analyzes the spoken words during board meetings and identifies subtle emotional undercurrents. For instance, it detects that a passive shareholder, let’s call her Anna, feels a loss of identity after the president remarks, “This part probably won’t interest you much, given that you’re just 6 International Holding Company Media Office, 27 February, 2024. https://www.mediaoffice.abudhabi/en/economy/artificial- intelligence-board-observer-appointed-by-international-holding-board-of-directors/ and Alissa Kole, A New Governance Paradigm is Necessary for AI-Powered Boards, in Harvard Law School Forum on Corporate Governance, April 21, 2024. https://corpgov.law.harvard.edu/2024/04/21/a-new-governance-paradigm-is-necessary-for-ai-powered- boards/#:~:text=Appointing%20AI%20board%20members%20has,members%20need%20to%20be%20considered. 7 International Holding Company Media Office, 27 February, 2024. https://www.mediaoffice.abudhabi/en/economy/artificial- intelligence-board-observer-appointed-by-international-holding-board-of-directors/ 8 Séverine Mulliez is a board member of the Association Familiale Mulliez and an expert in family business governance; some of the perspective offered here comes from her first-hand experience as a family member and director in a large, successful family business. -- 8 of 28 -- an artist.” AI doesn’t stop there; it offers actionable insights, suggesting the president could find this family member later at an opportune occasion, saying something like: “I want to express how much we value your participation and unique perspectives. I apologize for my earlier comment about you being ‘just an artist.’ What I meant was that I didn’t want to overwhelm you with financial details. However, I deeply respect your contributions and am happy to share any information transparently if you’re interested.” This simple gesture can make Anna feel valued and understood, restoring a sense of identity and inclusion. Without this intervention, the remark might have gone unnoticed, leaving an unresolved wound in Anna. What begins as a small misunderstanding escalates into a cycle of resentment, ultimately jeopardizing the family dynamic and the business’s potential. Yet Anna might bring invaluable insights, creative ideas, or a critical network of resources. AI’s ability to detect hurtful remarks, potential feelings, unexpressed psychological needs, and conflicts of interest, then provide recommendations, can be transformative, particularly during sensitive periods. (Léa Wang and Séverine Mulliez) 4. Challenges & risks As with any paradigm shift in business – which the evolution of AI certainly is – integrating this new technology into corporate governance systems and practices introduces a number of challenges and risks. These are challenges that must be navigated by all stakeholders, including executives and directors, investors, employees, regulators and others. In this section, we address these challenges and risks. Yes, the potential for AI integration into corporate governance systems is extraordinary and exciting; but such opportunity usually does come with significant risks if those opportunities are not managed properly. In this section, our contributors discuss some of these challenges: transparency issues, data bias, corporate culture challenges, regulatory concerns and ethical considerations. 4.1. Transparency & biases Transparency is one of the most important foundations of corporate governance. We rely on organizations being transparent in their financial reporting, their sustainability reporting, their organizational characteristics and their policies. This is true, regardless of which stakeholder is relying on the information. Investors need transparency to make buy-and-hold decisions (lack of transparency was a significant part of the Enron collapse, discussed in the Introduction); as scholars, transparent communication helps us create data and to understand what firms are doing and why. Of course, firms may not always want to be completely transparent; textual analysis and other techniques have helped increase implied transparency when the firm leans towards opacity. In theory, firms will begin moving towards greater voluntary transparency, knowing that we have the technology to infer their values. AI will only increase our power to interpret what organizations are doing, and it should only increase how organizations communicate with their various stakeholders. -- 9 of 28 -- Effective corporate governance is defined by the principles of transparency and accountability. To ensure that AI aligns with these principles, it must be characterized by transparency, that is explainability about the AI role in specific decisions, and accountability, that is, clear responsibility of who makes the final decision, the AI or the Board. (Christos Cabolis and Karl Schmedders) Transparency by the firm will no longer be an option. Employees, suppliers, regulators and – perhaps most importantly – investors have the tools to create their own stories based on what the firms do communicate. And what they choose to communicate will be subject to concerns of bias, both with respect to what is communicated and to how the information is generated. Institutional investors will increasingly demand that companies adopt sound AI policies, ensuring transparency, fairness, and ethical standards in their AI applications. Assessing how companies manage AI-driven risks, including algorithmic bias and data misuse, will become a critical part of stewardship strategies. (Jinsuk Choi) There are also risks related to data security and privacy, as the vast data processing capabilities of AI systems make them attractive targets for cyberattacks. Finally, bias embedded in AI algorithms, if left unchecked, may perpetuate systemic inequalities, undermining the very governance objectives they are intended to enhance. (Zhaozhao He, Mihail Miletkov, and Viktoriya Staneva) Perhaps we – as outsiders – will most readily appreciate how AI pierces the corporate veil and makes firms more transparent with respect to social and cultural issues. Jing Zhao and Richard Warr connect this to their work related to diversity on the board of directors. AI will absolutely help improve transparency in corporate governance information, in particular that associated with ESG or societal reporting that used to heavily rely on textual analysis software/expertise in the past. (Jing Zhao) AI at the simplest level will make corporate diversity policies and impacts more transparent. This could have both positive and negative effects depending on who is using the information. (Richard Warr) This final point is both fascinating and critical; transparency is generally thought of as a good thing, but the information that flows from transparency can be misinterpreted (by humans) or used to harm the firm or other stakeholders (by humans). It is our hope, perhaps naïve, that the human beings utilizing whatever information they take or create from AI methods do so appropriately; we can imagine many nefarious of inappropriate transparency. In the next two sections about culture and regulation, we address these concerns more. We conclude this section with a reminder that transparency is not just about what AI produces; firms need to be transparent in how they are using AI. We may still be in the early stages of this being necessary, but the need will increase over time. Stakeholders will want to know more and more about what the firms are doing; because if the firms are not clear and transparent, stakeholders will easily be able to figure out some things on their own. -- 10 of 28 -- Remember that AI projects are ongoing endeavors; establish a stopping point based on either the effort invested or the performance of the model. AI evolves over time, emphasizing the significance of transparency and explainability in its applications. AI generates valuable insights, but they need to be validated. This is where the ‘expert human touch’ comes in. (Małgorzata Smulowitz) 4.2. Organizational culture Culture is about growing, or cultivating, some desired outcome. Scholars have used many different definitions to explain ‘culture;’ perhaps our favorite is that culture is programmed behavior that distinguishes one group of people from another. The term ‘programmed’ acknowledges that culture is a choice and it can be managed. And, as organizations are collections of people, organizational culture comes to represent the behavior of its people that has been programmed by the organization itself. Corporate governance is – or should be – the programming function for any organization, whether that comes internally from the managers and directors overseeing day-to-day operations for externally from the C-suite and board of directors representing various stakeholder groups. Internally, AI represents both an opportunity and a threat. Employees know it is being used, and that it will continue to be used. Employees live the culture that the firm establishes. AI must be integrated in the institutional culture and values by supporting the organizational goals and objectives. (Christos Cabolis and Karl Schmedders) The organization should build a culture of trust to ensure that employees understand benefits that AI brings the firms and themselves at work. (Nguyen Thi Kim Oanh) A lot of process-driven low and mid-level jobs will simply disappear because AI can do it better. So, to me there are really two main questions. 1. Where are unique human skills still needed/jobs continue? 2. How do board members and senior executives adapt to this new world? (David Midgley) In June 2024, Volkswagen Group and Rivian entered into a $5 billion joint venture to partner on next-generation software for both groups’ vehicles. By November, Volkswagen’s software division Cariad, which was leading the partnership with Rivian, was in disarray, with budget overruns and an executive shake-up. Cariad employees want to development the technology in-house; Volkswagen’s senior leadership has other ideas. Effectively, Volkswagen was choosing to outsource development of vehicle software, which did not sit well with Cariad’s 6,000 employees. “We all learned about Rivian from the news,” said one senior Cariad engineer. After Volkswagen investing €12 billion over the past 4 years, many expect this to be the end of Cariad. Volkswagen leadership seem disappointed with the quality of the software technology that Cariad has produced, while Cariad employees blame Volkswagen’s overly -- 11 of 28 -- bureaucratic structure and inconsistent leadership strategies. “It was never a tech problem; it is a culture problem.”9 How board members and senior executives adapt to this new world will determine the organization’s culture. Can AI take on the role of a boss or CEO in a company in the future? This question lies at the heart of a key debate surrounding the balance between human decision- making and AI capabilities. While using AI in leadership roles could transform how companies are run, it also raises concerns about losing human creativity, empathy, and moral judgment. (Flora Huang) AI in corporate governance will become a fascinating study of artificial intelligence meeting human intelligence; the above question that Flora poses gets to the heart of this tension. We generally accept AI as an outstanding tool; but how much more can it be? How human can it be? Several of our contributors chose to answer this question, as it pertains to corporate governance and corporate governance research; their responses were unanimously aligned with those of Truls Erickson: Humans ought to be in charge of determining what is desirable, while generative AI may serve as a tool to enhance human creativity and help them to envision new possibilities. It basically means that it may be a good idea to use generative AI in various ways as a tool in visualizing new possibilities, while the judgment ought to be with the management team and the board. (Truls Erickson) What happens if we let the machines take over the high-level decision-making that has heretofore been the domain of high-level, highly-compensated human beings? Some of us do not want to find out. The fundamental root of various corporate governance problems lies in a natural weakness of human beings: greed. Currently, machines do not appear to have conflicts of interest. But once AI becomes more advanced, will this remain the case? There are already signs that AI exhibits certain human-like behaviors, such as refusing to follow instructions and generating hallucinations. It would be frightening if AI were to gain control - especially with malicious intent. (Horace Yeung) 4.3. Regulation Corporate governance has always been performed by human beings, for human beings. Perhaps that explains the consensus of contributors in this book who echo the above, that “the judgment ought to be with the management team and the board.” Yet those same scholars are also all enthusiastic about the potential benefits and opportunities associated with using AI in corporate governance. Thus, a line needs to be drawn. Boundaries need to be established. We will explore some of these boundaries from an ethical perspective in the following section; for 9 From The Financial Times, “Volkswagen’s $5b Rivian tie-up prompts dismay at software division,” September 9, 2024. https://www.ft.com/content/b861e949-76a2-4782-9c74-947dfb56b41a -- 12 of 28 -- now, we consider who might impose the necessary regulation – corporate leaders, investors, policymakers – and what that regulation might look like in different locations. We should also consider the corporate governance of "AI" as this transformative tool has been implemented more and more into corporate structure, operation, and management. Such as the ethical and legal considerations around using AI as well as who is held accountable for AI-generated issues, errors and biases. One notorious analogy in academia is the ‘hallucination’ problem. (Jing Zhao) The ‘hallucination’ problem in academia – and in AI – is the belief that falsehoods are fact. If AI guides a board of directors to make destructive decisions – possibly due to flawed pricing information, erroneous market metrics or incorrect legal interpretations – then who is responsible? According to ChatGPT, “the responsibility for relying on it primarily lies with the users or organizations deploying the AI system.” But the answer cannot be so clear and the legal implications may vary on a case-by-case basis. This is especially true in corporate governance dynamics; at the most basic level, a firm’s board of directors are appointed by the firm’s shareholders to represent their interests, in whatever ways they see best.10 Currently, it appears the liability issue is, at best, a very murky issue, one that will only become more clear over years of case law development. Thus, ironically, while transparency should increase through AI technology, responsibility for that transparency and the actions in leads to is still an evolving issue. Another critical challenge involves accountability. Distinguishing between decisions made by humans, hybrid systems, or AI alone could become increasingly difficult, leading to excessive trust in technological systems and jeopardizing human oversight. Compounding this issue is the presence of implicit biases in training data, which can distort outcomes, and the vulnerability of AI to cybersecurity threats, with potentially severe implications for corporate decision-making. (Cinzia Dessi) It may not be in a firm’s best interest to wait for clear and prescriptive regulation on AI usage and accountability. Waiting may lead to missed opportunities and sub-optimal performance. Firms can set their own policies and boundaries now; as Małgorzata Smulowitz mentioned earlier, “AI evolves over time” and “AI projects are ongoing endeavors.” When the environment changes, firm policies can change. For now, firms need to establish AI policies, procedures and goals: Our contrarian selves considered how existing corporate governance can shape the use of AI and, how this interaction will evolve over time. Given the justified concerns about biases in AI, because of the biases inherent in the material used to train large language models (LLMs), corporate structures will need to oversee the alignment of AI usage with the institutional goals. (Christos Cabolis and Karl Schmedders) 10 Will Directors’ and Officers’ Liability Insurance protect directors and officers using AI? We probably don’t know yet. -- 13 of 28 -- As this interaction does evolve over time, we are likely to see increased regulation on AI usage. This may add some needed clarity to some directors and officers, but the situation will likely always remain complex. The use of AI will potentially be regulated to different extends, ensuring responsible and non-exclusive innovation. In that case, especially in places like EU, corporations might have to report on their use of for instance generative AI tools or development in their business. (Agnes Dyvik Clark) We may need to develop regulatory standards for AI, fostering cross-border compliance, and ensuring trust and transparency in AI systems. Challenges such as AI interpretability, human-AI collaboration, skill development for directors, and ethical boundaries must also be addressed. (Arjya Majumdar) Even so, we may never have completely transparent and clear guidance on the legal implications of AI usage in corporate governance practices. We can think of international banks subject to different standards in different jurisdictions, varying internet policies, data privacy and others. For many readers and governance scholars, the dominant platform for data privacy is the European Union’s General Data Protection Regulation (GDPR). GDPR is focused on how organizations use personal data (similar to the Cambridge Analytica example). Yet, when those companies are using personal data in AI-driven algorithms and models, what are the risks? What are the responsibilities? If a bank customer is denied a loan based on AI-technology, does the bank have a responsibility to explain how that decision was made? Similarly, will a board of directors have (legal) responsibility to explain to shareholders how personal data (or other data) might have been used by AI to make decisions? The EU is also launching the EU AI Act in 202511; the goal of the AI Act is to guide EU organizations on how to use AI to benefit stakeholders. With respect to corporate governance, the Act has three priorities: enhance transparency, introduce risk assessments and include human oversight. These three priorities have been common themes throughout this book. The impacts of this Act remain to be seen, of course. The EU might like this Act to become a global standard for all businesses to follow; like GDPR, that might be unlikely to happen. Even as we gain some clarity on regional regulation, we may never have complete regulatory uniformity. Of course, that could be both good and bad. This technological revolution will challenge firm’s corporate governance systems, with or without clear regulator guidance and consistency. How firms integrate this technology and serve their stakeholders will determine their success at using AI systems. And, because corporate governance is about people, ultimate responsibility rests with the people in charge of these systems. How boards, board members and senior executives adapt to this new world will determine what happens next, for their firms and for broader economies. Outside tech companies, understanding of these new general-purpose technologies is limited. Do boards rely on external advisors? Do directors need to upskill themselves? 11 For details on the EU AI Act, see: https://artificialintelligenceact.eu/ -- 14 of 28 -- Do we need a Chief AI Officer? These and all the related questions about organizational transformations over the next 5-to-10 years need urgent attention. Or do we just let the creative forces of destruction work it out for us? (David Midgley) 4.4. Ethical considerations Ethical considerations have always been at the heart of corporate governance – and they always will. As Horace Yeung warned us above, “the fundamental root of various corporate governance problems lies in a natural weakness of human beings: greed.” For most of the past half-century, corporate governance research has explored ways to align different greedy people. We may call it incentive alignment or rational self-interest, but that’s the goal – to align greed. That is what we study. And, for us as corporate governance scholars, AI-related innovation creates a world of new research opportunities – both in how we perform our work and in the questions we ask. For us, AI is a revolutionary natural experiment. Now, as David Midgley questioned above, we will get to study how firms adapt to this new world; we will get to study how individual directors and leaders adapt to this new world. Concerning board composition, this revolution may require board members to obtain interdisciplinary knowledge, especially technological expertise in data analytics, machine learning, or AI algorithm to ensures that information and strategic insights are critically reviewed before decisions made. New guidelines and procedures on AI should be developed to ensure that AI usage is aligned with organizational core values and ethics. (Nguyen Thi Kim Oanh) Will firms create a board-level “AI committee?” Should they? In the U.S. in 2002, the Sarbanes-Oxley mandated that all listed firms have a ‘finance expert’ on the board of directors; will we see a similar mandate for boards to have an ‘AI expert’ on the board? The risk management committee should list and evaluate AI risks on long-term firm performance. AI adoption may cause new risks for the firms in terms of ethical considerations and transparency. As a result, board characteristics might change. (Nguyen Thi Kim Oanh) Our contributors were unanimous that the responsibility to ethical considerations rests with all stakeholders. Yes, the directors and officers represent the firm, which is the hub of this network of relationships; however, all parties must be accountable for their own actions, decisions and engagement. However, the extensive use of AI in corporate governance is not without potential drawbacks. A significant concern is that firms may increasingly rely on "hard" information while neglecting "soft" information, such as the nuanced insights gained through human interaction, which often contribute to sound decision-making. (Zhaozhao He, Mihail Miletkov, and Viktoriya Staneva) Public institutional investors must not view AI merely as a tool. They also bear responsibility for monitoring how investee companies leverage AI to achieve key -- 15 of 28 -- governance objectives, such as strengthening internal controls and improving stakeholder relationships. Understanding how AI can support both investors and corporations while mitigating unintended consequences will be essential. Effective governance in the AI era will require balancing technological efficiency with ethical accountability, ensuring that AI enhances long-term value creation rather than undermining it. (Jinsuk Choi) As we said above, how boards, board members and senior executives adapt to this new world will determine what happens next for their firms – but also for broader economies. It would be irresponsible for us to solely consider AI as a tool for long-term value creation for a singular firm; it also represents an opportunity (and a threat) for long-term value creation across entire economies, not just the elites entrusted with leading and governing large firms. Furthermore, the growing centrality of AI risks subordinating stakeholder trust to algorithms, overshadowing human strategic vision as well as ethical, moral, and social justice considerations is a major concern. Taking a macro perspective reveals implications that transcend corporate boundaries. The ripple effects of AI adoption extend to the nations where businesses operate, reshaping socio-economic dynamics. AI’s impact reaches beyond internal markets, influencing economies, redefining social structures, and creating new opportunities for development. Adopting this broader perspective helps us understand how AI-enhanced corporate governance can magnify its influence—not just on competitiveness but also on inequalities, labor market transformation, and resource redistribution. While technological innovation offers remarkable benefits, it also risks widening the gap between those who can adapt and those left behind. (Cinzia Dessi) And as societal standards evolve, public expectations for ethical leadership are rising. Corporate governance systems must consider not only what is legally allowed but also what is right. The Cambridge Analytica scandal underscores the importance of ethical decision- making in corporate governance. Cambridge collected data from millions of Facebook users without clear consent, employing psychological profiling techniques that were marketed as revolutionary but, by their own admission, achieved limited success. While their actions were arguably not illegal at the time, as no regulations explicitly prohibited them, they raised significant ethical concerns12. This case exemplifies the role of ethical frameworks in navigating such challenges. From a deontological perspective, which evaluates actions based on adherence to rules and duties, their behavior might be seen as acceptable since it complied with the existing legal framework. However, a virtue ethics perspective, which emphasizes character and moral values, highlights that their actions lacked integrity, transparency, and respect for individuals. Additionally, a utilitarian framework, which focuses on the greatest good for the greatest number, would question whether the potential harm to societal trust outweighed the company’s benefits. The public backlash against the perceived exploitation of personal data illustrates the consequences of ignoring broader ethical considerations. The 12 Freddy Gray, "His Dark Materials: A Year After the Downfall of Cambridge Analytica, Its Former Head is Ready to Talk," The Spectator, May 18, 2019, https://www.spectator.co.uk/article/his-dark-materials-16-may-2019/ -- 16 of 28 -- scandal serves as a reminder that ethical decision-making must align corporate practices with societal expectations, not just legal standards. 5. Amplifying research themes through AI Corporate governance is about how human beings related to one-another and how they make decisions. We care about AI because we are interested in how corporate governance systems utilize AI to serve their mission. The previous sections introduced a number of issues – transparency, culture, regulation, ethics, inequality – that need to be considered as organizations further integrate AI into their corporate governance systems. But, this is a book about advancing corporate governance research. Our job is to introduce current state-of-the-art research on corporate governance and to consider future directions and perspectives for new corporate governance research. And this is what makes the topic of AI integration in corporate governance so fascinating. Both corporate governance and AI integration are applied phenomena. We understand them better by empirically studying them. Our scholarly field thrives with natural experiments, or societal shocks that help us uncover how people relate to each other and how they make decisions. The AI revolution is just that – an external shock that we can use to better see how corporate governance systems function. And, even better, it’s unlike most other natural experiments – like new regulation or a tender offer – because it is both a choice and continuous. Organizations can choose to embrace AI in the corporate governance systems – or they can choose to not embrace AI. As scholars, we can observe those decisions to better understand what firms are doing and why they are doing it. Further, AI adoption includes a wide range of possibilities; some firms may use it merely for transcribing minutes of board meetings, while others use it to design strategic plans and even compose the board. These are all choices. These choices reveal preferences. And, to us as researchers, these choices become data. Considering that both corporate governance and AI integration are abstract, qualitative, human-centered dynamics, data is not always easy to create or college. This is the beautiful gift that this AI revolution has given us – and will likely continue to give us for many years to come. In Section 6, we imagine the AI-related topics that many of us will study in the future. For now, we want to highlight how AI integration might be impacting four novel areas of corporate governance research and practice: culture and international governance, venture governance, governance of family firms and governance of mission, impact and purpose. We want to briefly consider how AI integration is affecting the practice of corporate governance within each of these themes, which will help us imagine the future research directions and perspectives we will be studying in the future. 5.1. AI, culture and international corporate governance We introduced some opportunities and concerns related to AI, culture and international corporate governance in the earlier section on regulation. As we discussed in Part 1 on cultural dynamics, such seemingly external factors directly affect what firms do and how they are structured. We are already seeing different nations and geographic regions become AI -- 17 of 28 -- leaders, both in terms of development and adoption, while others are lagging. These differences will certainly be seen in boardrooms and governance systems. Although AI adoption is a global phenomenon, AI technologies are being developed in distinct entrepreneurial ecosystems and innovation “hotspots” (e.g., Montreal, Canada). Research is needed to understand how these ecosystems are governed in such a way that entrepreneurs are encouraged to pursue the responsible development of AI technologies. For instance, leaders of Montreal’s AI ecosystem have been very proactive in their attempts to shape the governance of AI innovation.13 But important questions remain about how collective action from diverse individuals, organizations, and entities in AI ecosystems is effectively governed to ensure that AI entrepreneurship is ethical and benefits society. (Philip Roundy) The governance of AI has become a common theme in our contributors comments; we cannot just consider how organizations are using AI in governance, but we must consider how they are governing the use of AI (and how external entities are governing AI use). A unique issue in governance research lies in the generational gap in technological aptitude, which often creates tension in decision-making power. Senior leaders, who typically hold significant influence in governance structures, may struggle to adapt to rapidly evolving AI tools. This resistance or unfamiliarity with technological advancements can hinder effective adoption and create disparities in governance practices. Understanding and addressing this gap is crucial for leveraging AI’s full potential while maintaining equitable and inclusive decision-making processes. We return to this issue on Section 6.3 as we connect AI adoption to the governance of family firms. Natural experiments can be wonderful for researchers, but they can be extremely difficult to navigate as business leaders and corporate governance systems. The point of introducing the generational gap now is to think about the social and cultural challenges that firms and economies will face as AI technology continues to advance. There will be adopters and laggards, possibly leading to winners and losers. Governing this will be difficult. A lot of process-driven low and mid-level jobs will simply disappear because AI can do it better. So, to me there are really two main questions. First, where are unique human skills still needed? And, second, how do board members and senior executives adapt to this new world? I think the answer to the first question is relatively straightforward, anywhere soft relationship skills are needed to work with others to build consensus, form coalitions, develop strategy, or anywhere high creativity is required, which might be more sector specific. I find the second question much harder to answer. (David Midgley) Much of what we have been considering in this paper is that second question – how do board members and executives adapt to this new world? As David suggests above, the answer to this will have real, human-level impacts. Jobs will change, lives will change. Making decisions 13 See, for example, https://montrealdeclaration-responsibleai.com/ and https://ia.quebec/en/governance. -- 18 of 28 -- that lead to these changes is not always comfortable; and, this will lead to challenges for regional and national policy makers, too. Over-reliance on AI could not only erode human critical and creative capabilities over time, but also standardize decision-making processes and foster growing dependency. Additionally, accelerated competition between AI-driven enterprises (and economies) may destabilize markets, leaving less adaptive organizations (and economies) at a disadvantage. (Cinzia Dessi) 5.2. AI and start-up & venture governance We now move from the macro-perspective to the micro-perspective, focusing on how the AI revolution will impact corporate governance with specific environments. And, in these next two subsections, we focus on two extremely different AI and governance environments. In the next subsection, we explore how AI is being used in the governance of family firms; without spoiling too much, it is safe to say that family firms might be late adopters. Does this mean they will be laggards? Not necessarily – it’s simply that the opportunities and challenges associated with incorporating AI into corporate governance manifest differently in different types of firms and different environments (which is why culture and context matter so much). We will discuss this more shortly. And if family firms are the late adopters, it is safe to say that start-up firms are the early adopters of integrating AI into corporate governance. In many cases, these firms are the ones creating the technology that we are discussing – of course they are going to be the champions. This familiarity and expertise doesn’t mean they are immune from the tensions and challenges that affect all types of firms. But they are also likely to be the first to enjoy the benefits. In the context of AI in venture governance, AI models can suggest the optimal structure for startups, depending on the phase of the startups are in, and what the founders want to achieve in specific timeframe. Ideally, this AI model can suggest board member based on the criteria provided by the founders, a more advanced version of board member directory. (Dennis Gan) Back in Part II of the book, we introduced two institutions that are driving the development of new ventures: entrepreneurial ecosystems and advisory boards. They both exist outside of the organization (informally or formally), but serve the purpose of advancing the start-up towards its mission. Importantly, they are dynamic institutions, evolving as the members and the start- up evolve; combined with the constantly evolving world of AI innovation, this again presents numerous opportunities and challenges for the firms. At the level of entrepreneurs and ventures, research is needed to explore how AI technologies can function as substitutes for (or complements to) the governance resources provided by entrepreneurial ecosystems. Entrepreneurs can look outside their organizations to the meta-organizational knowledge, networks, and norms of their local startup communities for the resources needed to govern their new and early-stage ventures. However, in some cases, AI technologies can provide an alternative to ecosystem resources and interactions (Roundy, 2022). (Philip Roundy) -- 19 of 28 -- Seeing how firms recognize these complements and alternatives will be fascinating to us as scholars. This will create numerous research opportunities for us, too. Research is needed to explore the governance implications of substituting AI technologies, which are based on algorithms rooted in calculative and formal rationality, for interactions in local entrepreneurial ecosystems, which are based on human connections rooted in substantive rationality (Lindebaum et al., 2020). AI and human interactions are fundamentally different and the corporate governance consequences for entrepreneurs choosing between them are not clear. (Philip Roundy) We can observe how entrepreneurs choose to engage in entrepreneurial ecosystems, but the composition and structure of those ecosystems is largely outside the start-up firm’s control. That is not true for advisory boards; start-up firms can choose (a) whether to utilize an advisory board, (b) what the advisory board looks like, and (c) what the advisory board does. AI’s integration into start-up advisory boards underscores its dual role as both a transformative tool for governance and something requiring careful oversight. This duality has implications for how start-ups – and organizations more broadly – navigate governance challenges. 1. AI as a Tool for Governance: AI can empower advisory boards by providing advanced analytics, predictive modelling, and streamlined decision-making processes. These capabilities enable advisory boards to offer precise, actionable insights. AI also enhances collaboration by automating meeting coordination, summarizing discussions, and tracking progress on strategic initiatives. 2. AI as an Object of Governance: Startups adopting AI technologies face ethical, legal, and operational challenges, such as mitigating algorithmic bias, ensuring compliance with data privacy regulations, and managing stakeholder concerns. Advisory boards are well-positioned to guide startups in addressing these challenges, fostering responsible AI practices that align with societal values and organizational goals. (Rod McNaughton) This last point is a critical one – AI integration is one specific case where advisory boards can provide significant benefit to their start-up firms, but how the firms utilize advisory boards is still a choice they get to make. And that makes it fascinating for us to observe the corporate governance implications of these choices. Advisory boards must balance leveraging AI for strategic advantage with maintaining ethical oversight. This includes advising startups on integrating AI responsibly into their business models while safeguarding transparency and accountability. (Rod McNaughton) What we are seeing in these perspectives – that are ostensibly uniquely about start-up firms – are the same issues we have seen throughout this paper: transparency, ethics, governance of AI, alignment with social and organizational goals. For many start-up firms, one common organizational goal is growth – transitioning from idea to unicorn (or at least to a larger and -- 20 of 28 -- more impactful organization). And that organizational transition will require the start-ups’ corporate governance systems to evolve, too. AI integration can help here, too. As startups grow and transition from advisory boards to formal governance structures, AI can be pivotal in institutionalizing knowledge, ensuring continuity, and supporting scalable decision-making. Advisory boards can help set the foundation for AI-enabled governance frameworks that endure beyond the startup phase. (Rod McNaughton) 5.3. AI and governance of family firms When we surveyed our 30+ authors on their perspective so the AI impact in their areas of expertise, the first response we received was clear and direct: “Currently, AI is not at all considered in the family business governance of the companies that we are familiar with.”14 But then as we began receiving more feedback – from that initial responder and others – we continued to see many of the themes and issues we have seen with start-up firms and with all other firms, too. At a basic level, family firms are using AI to enhance efficiency and the regular governance activities. In today’s world, family businesses that have adopted structured governance practices, including meetings and gatherings of all kinds, often touch on the topic of AI. The most common use they cite is, “We use AI to take notes.” Even for this straightforward function, concerns arise: Is the data secure? Who else can access it? Could there be leaks? While these questions linger, there’s a shared acknowledgment of AI’s efficiency in handling note-taking, especially during meetings involving many shareholders offering diverse perspectives. AI’s ability to summarize and organize discussions is undeniably impressive. Now, imagine this: AI goes beyond mere note-taking. It analyzes the spoken words during board meetings and identifies subtle emotional undercurrents. These contributions, though less visible, are vital to the family business’s legacy, harmony, and success. AI’s ability to detect hurtful remarks, potential feelings, unexpressed psychological needs, and conflicts of interest, then provide recommendations, can be transformative, particularly during sensitive periods like succession planning. Family businesses often face heightened uncertainty and competing interests during transitions. By recognizing divergences, understanding individual communication styles, and identifying fears and needs, AI can provide timely, actionable advice to improve dialogue and address conflicts early. (Léa Wang, and Séverine Mulliez) We saw the rest of this case study play out in Section 5.2. Rod McNaughton highlighted the distinction of start-ups using “AI as a Tool for Governance” versus “AI as an Object of Governance” previously. And we are seeing that with family firms, too. The opportunities and challenges are similar; so are the ethical, legal and operational issues. Léa and Séverine conclude their case study by imagining what might be next in the logistical integration of AI into corporate governance at family firms. 14 Attribution omitted intentionally. -- 21 of 28 -- If we push this idea one step further, we could envision integrating camera monitoring with note-taking, capturing and analyzing body gestures, hand gestures, facial expressions, and tones of voice that signal fear, anxiety, intimidation, control, or rejection, offering deeper insights. Imagine a robotic secretary observing the meetings, sharing her observations afterward, and being ready to answer questions and openly share concerns. The robot secretary would be open, free from personal interests, and non-judgmental—especially not pushy—leaving room for leaders to gauge their next steps themselves. Only insights would be offered. (Léa Wang and Séverine Mulliez) And that has been a theme throughout, too: the human leaders have the ultimate responsibility for making decisions and moving the organization towards its goals. This is, arguably, most relevant at family firms, where the goals may not only be about profit maximization and financial value-creation. As Małgorzata Smulowitz explains, many family firms choose to take a long-term, triple bottom-line perspective: people, planet and profit. In a separate article, Małgorzata and Peter Vogel study the Spandows family use of AI to enhance profitability, philanthropy and sustainability (Smulowitz and Vogel, 2024). The family is using AI to improve efficiency in disaster relief efforts, to decode honeybee dance patters, to map the Brazilian food system and to develop synthetic cancer data. That is, in this family business AI is being used exactly as it could be used at any other firm. AI is being used to create competitive advantages and to help the firm serve its triple bottom-line mission. In a competitive world for talent and innovation, staying ahead is crucial, and it is not getting any easier. By prioritizing the triple bottom line – considering social, environmental, and financial factors – businesses can stay relevant. By understanding and improving the business value, you can identify opportunities to improve and operate more sustainably. In the realm of corporate social value, progress is best made in small steps. One step forward may sometimes feel like taking two steps back, but it is a gradual and worthwhile process. Remember that AI projects are ongoing endeavors; establish a stopping point based on either the effort invested or the performance of the model. AI generates valuable insights, but they need to be validated. This is where the ‘expert human touch’ comes in. (Małgorzata Smulowitz and Peter Vogel, 2024) 5.4. AI and governance of social purpose, mission and impact Corporate governance is about how organizations provide a return on investment to suppliers of capital. That return can be financial; the return can be social or personal. Each organization gets to decide what the firm’s mission and purpose is; it is the corporate governance system’s responsibility to work towards that mission and purpose. The above example shows how AI can used to serve a social purpose, using business for good. Of course, AI doesn’t care what it’s being used for. That is up to use to determine; that is up to the humans in charge of corporate governance systems to determine. If a board prioritizes strengthening ESG initiatives, AI can serve as a complement to advance these efforts. We know that improvements in the S and G dimensions may yield tangible benefits for firms. Complementing the design and implementation of S -- 22 of 28 -- and G measures with AI can enhance the brainstorming, development, and implementation of policies decided by the executive team. (Christo Cabolis and Karl Schmedders) Once again, it all comes back to the executive team, the human beings responsible for making the leadership decisions. And, as we expand on the impact of AI in the governance of social purpose, mission and impact, we continue to see the recurring themes of transparency, data- driven decision-making and an explosion of information, for better and for worse. AI will absolutely help improve transparency in corporate governance information, in particular that associated with ESG or societal reporting that used to heavily rely on textual analysis software/expertise in the past. In addition, AI will transform corporate governance and its research by enabling corporations and boards to make more informed, data-driven decisions, more quickly. (Jing Zhao) This thinking is an extension of Jing’s Zhao’s study with Richard Warr on the impact of board diversity on firm innovation. Richard further provides a specific example related to his work with Jing Zhao on board diversity and firm innovation: AI allows very quick queries of diversity policies. For example, throwing ‘what are the pro-diversity policies implemented by Coca-Cola Company?’ into ChatGPT reveals a wealth of information. The follow-on question ‘what is the tangible impact of these policies for Coca-Cola?’ also reveals quite interesting stuff. (Richard Warr) As Richard reminded us earlier, just having more information is not always a good thing; ‘data’ can be a four-letter word: “This could have both positive and negative effects depending on who is seeking the information.” And that is perhaps the key point: AI technology can change the world of business and corporate governance, but whether those changes are positive or negative depends on how the technology is implemented and how the information it produces is used. Whether a firm is for-profit or for-mission, its purpose is some combination of profit, progress and prosperity. AI can help any firm better achieve that purpose. But human ownership and intelligence will ultimately be responsible for governing that purpose. Thus, the ‘expert human touch’ will always be the foundation of any corporate governance system. The challenge lies in designing a future where AI fosters both competitiveness and inclusivity. Policies and corporate strategies must transcend business boundaries, aligning innovation with societal equity and sustainability, ensuring that AI serves as a tool for progress rather than a source of human imbalance. (Cinzia Dessi) 6. Opportunities for future research The integration of Artificial Intelligence into corporate governance has begun to redefine how organizations operate and make strategic decisions. Boards and executives increasingly use AI tools for analytics, predictive modeling, and risk management, creating opportunities to -- 23 of 28 -- enhance decision-making and transparency. However, this technological shift brings new challenges, including algorithmic bias, data privacy concerns, and ethical accountability. And all of this will create – or demand – many new opportunities for research. Many of these opportunities have already been introduced in this paper and in prior papers. As we conclude this paper, we want to highlight some of these issues, both as they connect back to the 4 primary themes of this book and toward other themes and perspectives. As we have said a number of times, many of the issues related to AI integration in corporate governance are relevant to each of the governance themes in this book: cultural dynamics and perspectives, start-up and venture governance, governance of family firms and governance of social purpose, mission and impact. The specific opportunities and challenges might be slightly different depending on the context, but the issues of transparency, model bias, corporate culture, regulation and ethics matter to every type of firm. As we conclude this conversation, we first want to touch on a few perspectives on AI integration into corporate governance that can lead to novel research streams: ecosystem governance, decentralized autonomous organizations and institutional shareholder governance. These topics span the continuum of investor types. And they can help us think about the questions each of us should be asking as we push the frontiers of research on AI issues in corporate governance. Philip Roundy is currently pushing this frontier as it relates to entrepreneurial ecosystems and the intersection of AI, governance and entrepreneurs; see, for example, Roundy (2022) or Roundy and Asllani (2024). As we introduced above in Section 5.2, he sees a plethora of new research opportunities. We need to understand how entrepreneurial ecosystems are governed, such that participating entrepreneurs development and use AI technologies responsibly; we can imagine a free-rider problem with an ecosystem that negatively affects both the primary developers and all members. And, research is needed to understand the tradeoffs associated with substituting AI-driven connections for human-centered interactions within ecosystems. AI and human interactions are fundamentally different and the corporate governance consequences for entrepreneurs of choosing between them are not clear. This has been a theme throughout this paper, and it highlights our responsibility as scholars studying AI integration in corporate governance. Rod McNaughton is intrigued by an extreme version of this tradeoff, considering both the practical and scholarly potential of Decentralized Autonomous Organizations, or DAOs – which represent an explicit choice in favor of technology-focused decision-making and against human-centered governance. DAOs are an emerging governance model that use blockchain-based smart contracts to decentralize decision-making, offering an alternative to traditional hierarchical governance structures. DAOs could become an emerging governance model that challenge conventional assumptions. AI also intersects with DAOs in several ways, providing an opportunity to link the two: 1. Enhanced Decision-Making: AI systems can process real-time data to support decentralized decision-making within DAOs, providing insights that improve -- 24 of 28 -- transparency and efficiency. 2. Governance Automation: DAOs leverage smart contracts to automate governance tasks, and AI can enhance this automation by enabling adaptive decision-making that responds to dynamic environments. 3. Ethical and Regulatory Challenges: Similar to AI governance, DAOs face questions about accountability, fairness, and compliance. For example, who is responsible for decisions by an autonomous organization powered by AI? These are critical issues that require further exploration – just as they are in traditional governance models. Jinsuk Choi is also intrigued by the potential for AI and technological advancements to transform how governance is executed – but from a large firm, institutional investor perspective. AI is poised to reshape corporate governance in transformative ways, particularly in how institutional investors engage with and oversee their portfolio companies. For large asset owners like sovereign wealth funds, AI offers significant potential as a tool to enhance shareholder activities such as proxy voting analysis, corporate engagement, and stewardship of vast and diverse portfolios including global companies. AI-driven analytics can streamline the analysis of voting items and help identify governance risks and opportunities with greater precision. AI will serve as both a tool and a focal point for governance evolution. It challenges boards, management, and investors alike to rethink traditional practices and align technological innovation with sustainable, stakeholder-centered governance. For decades, most corporate governance systems have relied on a balance of democracy and dictatorship: the firm invites all shareholders and stakeholders to share voice their preferences, but ultimate decision-making rests with the board of directors, the board chair and the executive team. Will AI change that? DAOs aim to explicitly move away from this model, decentralizing both advice and decision-making. Large asset owners like sovereign wealth funds won’t want that; so do they resist innovation or do they avoid investing in firms that may be more suited to democratizing corporate governance? As Nguyen Thi Kim Oanh points out, “AI emphasizes organizational culture of high performance, high productivity, and high automation.” Can corporate governance systems establish organizational cultures that are amenable to the priorities before any negative consequences result from them? We saw earlier that Volkswagen has struggled to do so with its vehicle software development division. As we mentioned earlier, writing this paper has been more fun that most scholarly exercises are – because we have had the opportunity to listen to a diverse group of global experts, with vastly different perspectives on AI, on corporate governance and on firm mission. Throughout this paper, we have touched on many outstanding research questions, based on how we see AI impacting corporate governance in the future. We want to conclude this paper with a Top 15 list of outstanding research questions and opportunities for future research on the intersection -- 25 of 28 -- of AI and corporate governance. This list is non-exhausting and in no particular order; feel free to let us know if we omitted any essential questions that should have made this list. 1. How does AI integration fit into agency, stakeholder, stewardship, resource- dependency or any other theory of corporate governance? 2. How will existing corporate governance systems shape the use of AI and how this interaction will involve over time? 3. How will regulation of AI implementation evolve? Will the EU AI Act become a global standard? 4. Who will be held accountable for AI-based decision-making? 5. Will AI improve transparency between firms and stakeholders or will it create a ‘black box’ where stakeholders know even less about a firm’s decision-making? 6. How do boards of directors and executive leaders manage organizational culture while integrating AI-driven technology? 7. How collective action from diverse individuals, organizations, and entities in AI ecosystems govern AI implementation to ensure that it is ethical and benefits society? 8. How do entrepreneurs find the optimal balance between AI-driven systems and human-driven decision-making? 9. How can family firms and non-listed firms use AI to serve their mission, to manage communication and to enhance succession transitions? 10. How do institutional investors use AI to their advantage? 11. What happens to corporate ownership structures if decision-making moves away from employees and directors? Will employee and director stock ownership cease? 12. At the societal level, will AI worsen social and income inequalities? Will AI models impose historical data – e.g. board diversity structures – and make existing biases even worse? 13. What role will technology play in corporate governance in 10 years? Will Decentralized Autonomous Organizations become the next evolution of organizational structure, making transparency and accountability more opaque? 14. Can AI take on a voting role on the board? Can AI become the CEO? 15. How do board members and senior executives adapt to this new world? The integration of AI into corporate governance has begun to redefine how organizations operate and make strategic decisions. Boards and executives increasingly use AI tools for analytics, predictive modeling, and risk management, creating opportunities to enhance decision-making and transparency. However, as we have seen throughout this book, this technological shift brings new challenges, including algorithmic bias, data privacy concerns, and ethical concerns. AI integration is about creating new information – based on dynamic algorithms – so that corporate governance systems can make more effective decisions. Traditional, non-AI data models work really well when the world doesn’t change; AI-driven models can – but don’t always – work well when the world does change. And this is where the ‘expert human touch’ must take the lead. But, as we know, human beings can overreact to subjective factors, while -- 26 of 28 -- machines ignore subjective factors. In theory, AI over-emphasizes the short-term and under- emphasizes the long-term; all statistical and data-generation models do, but AI is data- generation on steroids, so the gaps can be even larger. The upside is that since AI output is dynamically generated by information, the gains in knowledge and accuracy are improving at exponential rates. To quote Richard Warr again, “this could have both positive and negative effects depending on who is using the information.” For us, as researchers working to better understand how corporate governance structures are utilizing and impacted by AI-driven technology, this has created – and will continue to create – significant opportunities to advance corporate governance research. Having common themes – transparency, decision-making, organizational culture, regulation, ethical considerations – across different contexts allows us to formalize theories and normative perspectives. From these foundations, we can then apply, test and explain the confluence of corporate governance and AI using case studies, qualitative storytelling and even rigorous quantitative analysis. The opportunities for corporations to advance their governance systems using AI-driven technologies are enormous – including transparency, increased efficiency, more effective decision-making, and better stakeholder engagement. But, there are real concerns, as well, as Cinzia Dessi reminds us: “the growing centrality of AI risks subordinating stakeholder trust to algorithms, overshadowing human strategic vision as well as ethical, moral, and social justice considerations.” The ‘expert human touch’ will always be – and must always be – the driving force behind any corporate governance system. Relative to many other streams of research, corporate governance research has uniquely focused on this human element. Corporate governance is about people. Our research attempts to understand their behaviors, incentives and decisions. And understanding these humans has always required sophisticated human intelligence from researchers. That will always be the case. But, the AI revolution has created a fascinating convergence of human intelligence and artificial intelligence – both for us as researchers and within the firms that we study. A few years ago, most of us could not imagine the impact that AI would be making on our lives today – and we could not imagine the impact that AI would be making within corporate boardrooms. Our job as researchers is to push our research into new directions and uncover new perspectives on corporate governance knowledge; and AI-driven technology has ignited enormous new opportunities for advancing our understanding of why businesses do what they do. The knowledge that we advance – at this intersection of human intelligence and artificial intelligence – will help shape what businesses do in the near future. We look forward to pushing the corporate governance research frontiers with all of you. -- 27 of 28 -- Reference: Jarrahi, M., Lutz, C., and Newlands, G. (2022) Artificial intelligence, human intelligence and hybrid intelligence based on mutual augmentation'. Big Data & Society, 9(2). Loughran, T,. and McDonald, B. (2011) When is a Liability not a Liability? Textual Analysis, Dictionaries, and 10-Ks, Journal of Finance, 66(1), pp. 35-65. Larson, B., Moser, C., Caza, A., Muehlfeld, K., and Colombo, L. (2024) From the editors: Critical thinking in the age of generative AI. Academy of Management Learning & Education, pp. 1–6. Roundy, P. (2022). Artificial intelligence and entrepreneurial ecosystems: Understanding the implications of algorithmic decision-making for startup communities. Journal of Ethics in Entrepreneurship and Technology, 2(1), 23-38. Roundy, P., & Asllani, A. (2024). Understanding AI innovation contexts: A review and content analysis of artificial intelligence and entrepreneurial ecosystems research. Industrial Management & Data Systems, forthcoming. Smulowitz, M., and Vogel, P. (2024). The Spandows: Pioneering AI in family philanthropy and sustainable business. In The Routledge Handbook of Artificial Intelligence and Philanthropy, pp. 275-286. Routledge. -- 28 of 28 --