AI: exploiting uncertainty and redefine organizational governance

We live in “an era of mass extinction”. According to tech-entrepreneur Thomas Siebel close to half the companies in existence today are likely to disappear within ten years. Generative AI offerings such as ChatGPT4 will only accelerate that trend. Boards are now challenged to re-evaluate their purpose, relevance, and capabilities to survive in the age of AI.

Historically, every leap in technology eventually challenged organizational designs.  Current trends are tech-empowering frontline workers to self-organize, consequently creating more flattened and distributed organizational structures.  However, given the impending AI-fueled transformation, the majority of ‘knowledge jobs’ at the frontlines are likely to be either automated or disappear altogether within the next 10-15 years. In the absence of deep AI understanding to supervise, manage, and complement AI technologies, no ‘knowledge worker’ – will survive. The CEO and the board of directors are equally challenged.

We tend to assume that innovation and transformation are all about CEO leadership when it is increasingly a matter of organizational governance. Management cannot do it alone. To take an extreme example, Google’s executives were so flummoxed by ChatGPT’s threat to its $150 billion search business that they called back the company’s founders. Directors must now work with owners and executives to re-define corporate governance for the AI age.

What Boards Can Do

As AI decouples prediction from judgment, human judgement is becoming more important, and companies will need to expand the ‘smarter human, smarter machine’ interface leveraging the more distributed decision-power towards the frontlines. The board must redefine its fiduciary duties accordingly, and Vijay Govindarajan’s Three Box Solution – preservation, destruction, and growth – may lead the way.

(1) Preservation – Integrate AI expertise at the board level. The board’s traditional supervising and advising responsibilities must now take on the ethical, regulatory, and safety reviews of algorithms. As algorithms become more powerful (and potentially damaging), boards must increase their expertise in this area, perhaps through a dedicated AI-dedicated board committee with access to specialist knowledge outside the firm. Overall, we should expect more government regulations in this area, broadening the board’s compliance mandate. Today, boards are not equipped to meet this challenge. Often, cognitive diversity and critical reflection capabilities in these future topics are missing.

(2) Destruction – Destroy structures and conventions that bury uncertainty. Organizational scaffolding such as catalogues of rules, hierarchical decision-making, best practices and, possibly CEO dominance, for example, all may have made sense in earlier decades, but now these can breed inertia and prevent the organization from embracing the uncertainty companies need to thrive. Uncertainty is the fuel to optimize prediction models. The faster prediction models can be tested and improved under uncertainty, the better their predictions, and the less uncertain a firm’s future. Uncertainty thus becomes the source of collaborative advantage. Indeed, embracing uncertainty requires destroying an uncertainty-resistant mindset that still rules the boardroom and corporate leaders alike.

            (3) Growth – Mobilize for massive up-skilling with fluid top-down/bottom-up connectivity. First, boards can promote AI-driven up-skilling across the entire organization, without waiting for the CEO to act, partly by pressing for dedicated HR staff to carry out the training and raise awareness for the threats and opportunities AI represents to individual tasks and jobs across the firm. Second, as AI is only as good as the underlying data is relevant, boards must review business models and strategies in view of optimizing access to good data. Business model innovation will now also serve the purpose of accessing new and relevant data.

Third, the board must insist that management articulates and communicates strategy with an eye to prediction models to generatively minimize error through action-based (and increasingly machine-generated) feedback loops.  The AI-fueled company requires leaders to develop a new language to AI-empower frontliners, learn to think probabilistically, and embrace bottom-up data to challenge predictions. In fact, boards must assess a CEO’s AI-readiness and judgement faculty by the person’s active enthusiasm for being falsified to generatively improve prediction models.

Promising Areas

Only a few boards have started considering and working on any of these dimensions, and we have no examples yet that can be comprehensively profiled. One promising practice that may be leveraged to impact all three dimensions are so-called ‘shadow boards’. Young, talented high-potentials, below the senior level, hold meetings with and without the regular board, using the same agenda and defining future agendas. 

Helpful are shadow boards, as IMD’s Jennifer Jordan has analyzed Prada, Gucci and AccorHotels. Not only does the board learn the perspective of younger employees (who often know more about emerging areas and trends than their elders), but the practice can identify opportunities and threats early on and make them available for top-down decision-judgement. But, more importantly, shadow boards can now serve as a powerful tool closer to the frontlines for mobilizing AI up-skilling, reviewing future data strategies, and cultivating language and thinking in terms of prediction models to better harness the predictive learning power of AI.

The greater the opportunity for systemic impact, the greater the board’s role is to ensure a holistic approach and continuity beyond the CEO. In the U.S., for example, several advanced hospitals such as Mayo and Cleveland Clinic are discovering that task-specific AI empowerment in radiology analytics, for example, has repercussions far beyond the radiology department – affecting their entire value chain.

Singapore’s DBS, for example, invested heavily in AI and has already moved from laggard to Southeast Asia’s leading bank.  It now employs more data scientists than bankers. Its Group-CEO, Piyush Gupta, has strong roots in technology. Though, as important as a tech-gifted CEO will increasingly be, it is uncertain how DBS’ AI-fueled re-invention continues and accelerates beyond Gupta without a board taking the steps outlined above.

The Challenge with Boards

Why aren’t boards further along in this?  Do boards themselves really want to step up? Do CEOs have any interest in boards to step up? These are pertinent questions, and the answers will vary individually and by company. The more senior and broadly active the directors, the more likely they are to rely on generic compliance formulas and best practices to assess management. For all the talk about the “revolt in the boardroom,” directors have actually become collectively – maybe not individually – estranged from innovation.  In their inaction, they have inadvertently reinforced the CEO’s dominance and potentially fostered inertia. On the CEO front, truly great leaders want a board that can make their performance stronger and more sustainable. But CEOs have no time and attention left to ‘up-skill’ the board.

The future of AI governance is holistic. Fundamentally, it is a culture change that cannot be left to executive leadership alone. It starts with owners and the board. Private equity sponsored governance models and to some degree family-controlled businesses have turned the board into a productive force long time ago. But the publicly listed corporation appears more vulnerable than ever and its eclipse is more real today. Companies will need directors that go beyond their fiduciary duty of oversight and are empowered and skilled to support the AI-driven transformation, actively and holistically.

Dr. Bijan Khezri is Chairman of KCRI (Khezri Capital Research International) AG (Switzerland) and Group-CEO of Marquard Group (Switzerland). He is the book author of ‘Governing Continuous Transformation: Re-framing the Strategy-Governance Discussion’ (Springer, 2022) and ‘Generation Dubai: Exit, Voice, and Loyalty’ (2018).

 Prof. Dr. Oliver Gassmann is Chairman of the Institute of Technology Management and of the Global Center for Entrepreneurship and Innovation at the University of St. Gallen. He is one of the most cited innovation scholars, book author of ‘The Business Model Navigator’ and serves on several boards.