The spectre of boards hosting their own artificial intelligence directors seems to materialise with the annoying regularity of pop-up ads. But how would non-executives cope if their chair proposed bringing one on board?
Academics in Australia have some timely advice: keep the AI fixed on a very narrow set of decisions and away from those that need human consideration.
In a research paper released this week, law academics at the University of New England in Adelaide warn boards that AI directors should be restricted to “administrative” decisions and kept a distance from “judgment tasks”.
It’s also important, they write, that independent directors will still be necessary to monitor AI decisions. That way stakeholders still have someone to make accountable for AI’s “acts and omissions”.
“There is no doubt,” write Mia Mahmudur Rahman and Pritam Dey, “that the impact of AI on corporate governance is profound.
“AI is redefining the concept of rights and duties of various actors in a company, affecting capital markets and consequently our ideas of companies. Hence, today’s directors need to be proactive in learning to work with AI.”
In corporate governance circles, there has been frequent discussion of AI, or robots, taking a greater role in boardroom decisions. Narratives have long mentioned Hong Kong venture capital firm, Deep Knowledge Ventures, as appointing an AI algorithm to its boardroom as early as 2014. Though much cited, the story has been dismissed by those in the know as “publicity hype”.
Judgment day
Despite the lack of active examples, the debate has moved on from whether AI should be used in boardrooms to whether it would be legal to use AI.
This line of inquiry focused on potential changes to the law because the UK Companies Act currently asks directors to act with “independent judgement” and “reasonable care, skill and diligence”.
Elsewhere, the worry is legal provisions that directors should be “legal persons”. However, some believe it possible to sidestep that concern by redrafting company constitutions.
In the latest research, the writers worry how AI might affect the role of directors.
They resolve this by suggesting AI should be used only for making administrative decisions—those that are routine and involve scheduling and reporting. What AI should not be engaged in, they argue, are judgment tasks: “managing organisational structure and company culture, talent management and shareholder and stakeholder management”.
Oversight by independent board members will also still be essential. “They are already better suited to keep an impartial check on inside directors. Hence, it will be easier for them to keep a check on a company director’s ability to make decisions based on AI.”
Whether AI robots will find their way onto the boards of the world’s largest companies remains unknown. While it may spark titillation to ponder robot directors, the current tech landscape would suggest we may be some way off. After all, companies continue to grapple with the governance of AI more generally, let alone its role in making board-level decisions.
That said, it’s as well we work through the governance options. Technology never rests and AI is on standby for a boardroom seat.