Can Boards of Directors govern AI?
By Mario Lara, Director of Esade Madrid and the Center for Corporate Governance.
NéstorSalcedo, Researcher at Esade's Center for Corporate Governance.
The advance of information technology (IT), and in particular artificial intelligence (AI), has permeated every corner of our society, transforming the way companies govern themselves and operate their businesses, and, consequently, their boards of directors face the different contents of their agendas including ESG issues.
The recent Bletchley Declaration signed in the UK in early November by more than 28 countries to promote the safe and responsible development of AI, centered on people and for the benefit of the global community (AI Safety Summit, 2023) is probably a good indication of the sensitivity that exists today to support efforts for the ethical development of AI. However, the recent crisis of OpenAI, creator of ChatGPT, in mid-November and the fracture generated in its corporate governance, with a strong background debate on the approach and future efforts for the development of AI, have raised alarm bells about whether we will be able to match the development of this technology with an ethical and responsible approach. As a sign of the strong debate within the company and society, in just one weekend, pressure from shareholders and investors and the announcement of voluntary resignation of all employees led to the resignation and restructuring of the entire Board and the return, after their termination, of both the President and CEO of the company (Roose, 2023). These events bring to the table the debate on, how can boards of directors strategically adopt and govern IT as complex and specialized as AI, and how to incorporate it into their agenda and into their decision-making and oversight processes while ensuring an ethical view of its use?
Answering this question requires a board of directors, on the one hand, to have an assessment of the impact of AI on the value chain of the industry in which the company operates, to identify the risks that can generate a disruption of such high impact on its operating or business model, and to have a good understanding of how the management team approaches the management of IT and in particular of generative AI; but on the other hand, it requires that the board itself, with the appropriate level of technological competence, understand what use they can make of generative AI to support their own work in fulfillment of their fiduciary duties.
A recent report on technology and corporate governance by Xvalue, Capgemini and Telefónica (Estévez and Morgado, 2023) indicates that 90% of Spanish company directors consider IT as a key factor for the future of business strategy, and 71% believe that AI will play a fundamental role in their company. However, 87% believe that boards should improve their IT profile and knowledge. A study by Esade CGC in collaboration with Diligent (Lara et al., 2022) found that, out of 5295 listed companies from 50 countries in North America, Europe, Asia and Oceania, only 8% of directors had IT expertise. These data show a great opportunity to improve the effectiveness of Boards, but also pose significant risks, ethical dilemmas and challenges that require commitments and actions from the Boards themselves.
Risk-based dilemmas and challenges
The adoption and use of AI poses dilemmas and challenges in corporate governance that require Boards to have a specific categorization of their risks (Esfera Consejeros, 2023; PwC and OdiseIA, 2022). Risks ranging from data protection and security to fairness, transparency and business performance measurement, all fueling the need to develop new regulations to safeguard from an ethical perspective the rights of different stakeholders, especially employees, suppliers and customers, and impacting the management of global issues such as, for example, the disruption of supply chains or the management of natural resources.
Likewise, the development of large-scale and generative AI models, such as ChatGPT, Midjourney, or DALL-E 2, have demonstrated innovative capabilities, but are not free of business risks due to the inherent opacity of their algorithms. Several universities, such as Harvard, Oxford, or Stanford, warn about the possibility that they may incorporate biases that may imply serious technological, reputational, and compliance risks, highlighting the complex ethical dilemmas associated with their development and implementation.
Another challenge identified is the operational risks of integrating AI into business processes. While AI can deliver performance improvements and generate operational efficiencies, its adoption poses enormous ethical challenges because of impacts on key decisions in people management or resource allocation, which require special oversight by boards of directors
The implementation of AI helps to bring areas such as IT investments and talent development and management that traditionally have not had much attention on their agendas to the Council debate. The need to execute very high investment IT projects forces Councils to monitor the training and availability of specialized talent, as well as their attraction and retention strategies. In addition, these technologies create new jobs, but at the same time they project the disappearance of traditional jobs, awakening reputational risks due to uncertainty in the workforces because of potential restructuring processes due to obsolescence, and raising the need to address massive reskilling processes that minimize their social impact (S).
In addition, the rapid development and adoption of new technologies such as AI can exacerbate risks to the environment. Increased energy consumption due to the high requirements of data centers and the need to train and operate AI models may increase the volume of carbon emissions. The generation of new e-waste, the intensification of natural resource use and the impact on biodiversity are other risks (E) to be controlled by councils.
These challenges and the fact that AI recreates the biases with which its models were created, invents data, extrapolates facts or words, and its use can contribute to leaking sensitive data to the market, require Boards of Directors to address critical questions: How to ensure that AI models developed or adopted by the company are ethical, ensure the security and privacy of information, or do not perpetuate bias or discrimination? How can we prevent the use of AI from causing us to make serious errors of judgment and leading to unfair decisions? What is the impact they can have on our workforce? How can we ensure the necessary transparency, understanding and explainability of decisions made by AI? To what level should or can we delegate AI-based decisions? Is it feasible and under what conditions to make AI tools available for board decision-making processes? Governance (G) around AI is critical to avoid these risks, ensure proper accountability and maximize its long-term benefits.
Commitments from the Boards of Directors
Boards can support the adoption of responsible AI, (van der Mullen, 2021), highlighting as some relevant commitments:
- Using AI in defining corporate purpose: This involves using AI to analyze varied data from multiple sources, better understand stakeholder expectations and contribute to the redefinition of corporate purpose, generating some clear advantages such as reducing information asymmetries between governance and management, as well as strengthening reputational legitimacy.
- Identifying and prioritizing ESG-focused IT investments (i.e., AI): This will translate into using AI to improve internal processes, reduce environmental footprint or meet stakeholder expectations on sustainability. It is essential for Boards to be aware of changing societal expectations in ESG, not only to comply with regulation, but to anticipate and proactively respond to these demands. AI can help to better monitor and analyze these trends.
- Integrate AI into governance: The adoption of responsible AI can be driven by formalizing its inclusion on the Board agenda or in a dedicated IT/IA committee. This committee can address AI governance in a holistic manner, identifying how it can be applied in areas such as diversity, equity and inclusion (DEI), talent management, regulatory compliance or investor communication. In addition, from a long-term perspective, they can evaluate and drive IT investments, prioritizing technologies that best contribute to sustainability challenges.
- Improve non-financial communication to the market: AI can help not only to understand stakeholders ' needs but also improve efficient and transparent reporting, strengthening confidence in the company's ESG commitments.
AI Governance Structure Models
Finally, we would like to highlight how some studies (Forrest, et al. 2022), delve into different models of Council participation in IT governance with an emphasis on AI, to effectively address this strategic challenge.
- Regular and full participation of the board of directors: This model is suited to companies where technology is vital to their business. It requires all board members to understand the technology implications and actively participate in strategic discussions and decisions. It is not exclusive of an advisory board to general management to evaluate specific areas and support in critical incidents.
- Permanent technology committees: especially for companies that are not technology-based. These committees help connect corporate strategy with digital, prioritizing technological decisions and overseeing aspects such as cybersecurity.
- Temporary commissions: Useful in times of major transformations or high IT investment. They provide specialized monitoring and validation at key moments, such as transitions to the cloud or major cyber events.
- Informal participation of the board of directors: Most appropriate when only one or two directors have extensive technical expertise and management needs specific advice on strategic IT/IA issues.
These models can coexist or be combined and their choice will depend on the specific situation and needs of each company.
Ultimately, the adoption of AI by Boards involves significant dilemmas and challenges. The key lies in finding a balance between addressing its challenges, mitigating its inherent risks, and maximizing the benefits across all dimensions of responsible AI business management to ensure a positive long-term impact for the company.
References
AI Safety Summit (November 1, 2023). The Bletchley Declaration by Countries Attending the AI Safety Summit, 1-2 November 2023. gov.uk. URL: https://www.gov.uk/government/publications/ai-safety-summit-2023-the-bletchley-declaration/the-bletchley-declaration-by-countries-attending-the-ai-safety-summit-1-2-november-2023
Estévez, R. and Morgado, J. (May 2023). La tecnología desde los consejos de administración. Esade Corporate Governance Center. URL: https://www.esade.edu/faculty-research/es/centro-de-gobierno-corporativo/publication/la-tecnologia-desde-los-consejos-de-administracion
Esfera Consejeros (2023). Risks and opportunities in the use of AI in the enterprise. Digital Transformation Series. InFocus. Madrid: Institute of Internal Auditors.
Forrest, W., Li, S., Tamburro, I. and van Kuiken, S. (September 15, 2022). How effective boards approach technology governance. McKinsey Digital. URL: https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/how-effective-boards-approach-technology-governance
Lara, M. Aguilera, R., Infantes, P., Nascimiento, K., Arrarte, R., Frimptong, E. (2022). Environmental (E) and Socially (S) friendly boards: Linking board effectiveness attributes with E and S performance. Esade Corporate Governance Center. URL: https://www.esade.edu/itemsweb/wi/research/cgc/CGC_Informe_Environmental_and_Socially_friendly_boards_Linking_board_effectiveness_attributes_with_E_and_S_performance.pdf
PwC and OdiseIA (2022). Guia de buenas prácticas en el uso de la inteligencia artificial ética. Madrid: PwC. https://www.pwc.es/es/publicaciones/tecnologia/assets/guia-buenas-practicas-uso-inteligencia-artificial-pwc-odiseia.pdf
Roose, K. (November 21, 2023). We explain the chaos at OpenAI. The New York Times. URL: https://www.nytimes.com/es/2023/11/21/espanol/sam-altman-openai.html
van der Mullen, R. (November 19, 2021). Looking to create an ESG program? Keep these three tips in mind. Gartner. URL: https://www.gartner.es/es/articulos/buscas-crear-un-programa-esg-ten-en-cuenta-estos-consejos
Mario Lara, Director of Esade Madrid and Esade's Corporate Governance Center
Mario Lara is Director of Esade Madrid and of the Esade Center for Corporate Governance. He is Academic Director of the Program for Directors at Esade and of the ESG program for directors. A specialist in the third sector and corporate governance, he has thirty years of experience in advising large national and international groups, as well as listed companies.
He has spent most of his professional career at PwC, where he joined in 1997 and was promoted to partner in 2004. He is an independent director and member of the appointments, remuneration and ESG committee of the board of directors of the Prim group. He has been Senior Advisor for Spain at Georgesondes since 2019.
Mario holds a degree in Law and Corporate Counsel from Universidad Pontificia Comillas ICAI-ICADE. He holds an executive MBA from the University of Houston, completed the Corporate Strategy program for Board Members at Esade and completed the Senior Management Development program at IE Business School.
His areas of focus include corporate governance, the nomination and compensation committee, ESG issues and transition to the board.
Néstor Salcedo, Researcher at Esade's Corporate Governance Center
Nestor U. Salcedo is a postdoctoral researcher at ESADE's Center for Corporate Governance, where his recent studies focus on the transition of executives to directors, the dynamics of board committees, the relationship between corporate governance and information technologies, and comparative corporate governance in Ibero-America.
He is also an academic member of the European Corporate Governance Institute (ECGI), the International Corporate Governance Society (ICGS), and an adjunct professor at ESAN Graduate School of Business.
He is also a member of the Editorial Boards of the Journal of Economics and Development (JED), Corporate Governance and Organizational Behavior Review (CGOBR) and Journal of Cases on Information Technology (JCIT). He has served as Editor-in-Chief of the Journal of Economics, Finance and Administrative Science - JEFAS (2021-2023).
D. in Economics, Management and Organization from the Autonomous University of Barcelona, MRes. and MBA from ESAN University, and BSc. in Architecture from the National University of Engineering of Peru.
