Happy New Year!
Among the top issues on board agendas in 2024, generative AI stands out as the most urgent and potentially vexing.
AI technology is rapidly becoming more sophisticated and accessible, ushering companies into unfamiliar territory and raising questions about data security, the future of work, and—considering who is largely driving its development and adoption—AI’s potential impact on underrepresented groups. It’s clear that generative AI will bring disruptions to companies in every sector, and it’s fair to ask whether boards are equipped to deal with its implications.
Chris Smith, chief strategy officer at consulting firm Grant Thornton, says he believes boards can meet the moment, but worries that many are starting from the wrong place. On one hand, boards are built to weather big threats and “be the reliable constant.” However, he hasn’t been encouraged by the governance response to generative AI that he’s seen so far.
In the fall, Grant Thornton polled a cross section of board members during the National Association of Corporate Directors’ global conference, asking how many of the 1,000 or so boards represented had held a discussion with management about their respective philosophical and cultural beliefs around AI. It found that less than 30% had done so. (Just a few weeks after that NACD event, the same question about alignment between boards and management would be one of the key issues to cause a now-infamous rift between OpenAI CEO Sam Altman and the company’s former board.)
The poll results indicate what this year should be about for boards, according to Smith: getting back to basics. Boards should be asking essential questions like: Does the board and management see AI as a force for good, a danger to world order, or something in between? Do we need to use AI to augment employee efficiency and effectiveness? And should we even use it at all? “Just because you can, doesn't mean you should,” he says.
Most importantly, a board of directors has a responsibility to understand AI’s impact on all stakeholders, including investors, customers, employees, suppliers, and the community, Smith adds. An AI-enabled product or workflow enhancement “might be right for investors, and might be the right thing for the customers, because it could drive down prices,” he says. But companies ought to ask: “What is it doing to the employee base and every community that houses those employees and future employees?” Starting with this conversation, he argues, will allow boards to ask the right questions.
Smith and Peter Gleason, CEO of the National Association of Corporate Directors (NACD), both suggest that boards first focus on AI as a company risk, rather than an opportunity. Allow management to come to the board with the possible upside to adopting AI, says Gleason, but boards should also look at the worst-case scenarios. For example, consider that a data breach or a malicious deepfake video of a CEO can quickly wipe out a company’s hard-earned reputation. “This is probably one of the most powerful things that has hit industry since the iPhone, and there are no guidelines,” he points out. “We also gave this tool to bad guys with no limitations on it.”
Boards should closely track a company’s intentions and plans for AI algorithms, says Smith, and continually verify that its AI tools are behaving the way they expect. Boards may need to make AI a topic of every board meeting if that’s what it takes to stay on top of a company’s AI deployment. Today, he sees companies allowing “rampant use of AI in a very decentralized way at all levels of the company,” he observes, and that makes him nervous.
Lila MacLellan
lila.maclellan@fortune.com
@lilamaclellan
