Two weeks ago, I attended the Collision conference in Toronto, a major tech summit where generative artificial intelligence dominated the conversation.

In a keynote talk, the grandfather of A.I. himself, Geoffrey Hinton, took to the stage to repeat warnings about A.I. and the potentially catastrophic problems it might generate. He called for engineers to slow down and put guardrails in place before A.I. becomes more intelligent than its makers. “I think the people developing it should be encouraged to put a lot of work into understanding how it might go wrong, understanding how it might try and take control away,” he said.

On the whole, however, A.I. enthusiasm saturated the event, just like it has everywhere else. (“Generative A.I. is churning out unicorns like it’s 2021,” a research firm declared in a recent market snapshot, noting that 13 A.I. companies are now valued at over $1 billion each.)

Large companies and their boards have begun to move beyond the hype to a place where they’re thinking more critically about A.I., according to Jay Persaud, a partner at global professional services firm EY, where he’s on the hunt for innovative startup technologies, including A.I., that can solve existing problems for EY and its clients.  

On the one hand, boards hear that A.I. is the “most important invention since the internet” and an absolute must-have, Persaud says. But generative A.I. software relies on large language models that must be customized with a company’s data to be useful. That opens the business to significant risks that boards are responsible for monitoring.

“Boards are also looking to management to come to them with a balanced discussion,” Persaud said, including a plan for safeguarding data. “Boards are saying, ‘Yes, we’re going to use [A.I.], and we’re going to get really good at it…But how do we protect the crown jewels? How do we keep ourselves out of trouble?’”

Peter Gleason, CEO of the National Association of Corporate Directors (NACD), tells Fortune that his organization views A.I. as transformative. But he also sees boards “taking a contemplative, inquisitive approach in assessing company and industry implications.”

A recent NACD pulse survey shared with Fortune shows why caution may be warranted: Only 10% of directors surveyed said their management teams are very or extremely proficient with A.I. issues.

Lila MacLellan
lila.maclellan@fortune.com
@lilamaclellan

Noted

 “Girls looked at us and said, ‘You’re not relevant.’”

—Richard Dickson, Mattel’s president and COO, describes the company’s tough realization about Barbie a few years ago in Bloomberg Businessweek.

On the Agenda

👓 KPMG just published a detailed advisory memo explaining how boards can set up a framework for assessing A.I.’s risks. It’s a daunting read, warning that possible A.I. threats include “data poisoning, model poisoning, back doors, insider threats, and other ways that attackers might damage the company’s decision-making systems.”

🎧 A.I. appears to be fueling a comeback for Big Tech. The Big Technology Podcast asks Wall Street Journal reporter Eliot Brown whether the numbers make sense.

📖 A survey by executive search firm Russell Reynolds found that only 37% of boards and leaders feel prepared to deal with geopolitical disruptions. In a new commentary, the firm argues that to combat that shortcoming, more boards must balance their teams of specialized experts with “networked generalists,” defined as leaders with “specific personal and professional experiences” that transcend a single sector.

In Brief

- Shares of Icahn Enterprises, Carl Icahn’s publicly traded holding company, surged on news that the activist investor has a new agreement that separates his personal loans from the share price of his company stock. Former risks tied to the loans were recently highlighted by the short-seller Hindenburg Research, whose attack on Icahn resulted in severe losses for the legendary investor.

- Randall Stephenson, former CEO of AT&T, resigned from PGA’s advisory board over its proposed merger with the Saudi-backed LIV Golf. Stephenson said the board had been left out of the loop as the company arranged the deal, despite the considerable ethical questions it raised.

- The Body Shop created a Gen Z advisory board to bring the voice of a younger generation to the business. Getting the project off the ground took two tries.

- Meta’s massive and mostly well-reviewed “community forums” system of democratizing the governance of metaverse issues will soon be rebooted to take on emerging A.I. concerns. The last initiative involved over 6,000 Meta employees across dozens of countries who tackled meaty topics in small discussion groups.

- A small percentage of board members feel they have the expertise to challenge management about sustainability matters, according to a new report from Heidrick & Struggles, INSEAD, and BCG. Surprisingly, two-thirds of directors surveyed say sustainability has little effect on their companies’ finances today.

The Long Read

Assume you’ll see a lot of Pantone 219 C, Barbie’s specific hue of pink, in the next few weeks as we approach the opening of Greta Gerwig’s much-hyped film. If you're wondering how Mattel managed to license a clever, of-the-moment movie about the blonde doll, you’ll want to read The New Yorker’s feature about the creative and strategic business decisions behind the movie. Mattel's first step, journalist Alex Barasch reports, was to hire Robbie Brenner, a Miramax veteran, to lead a new division: Mattel Films. 

“At the start of the ‘Barbie’ process, Gerwig decided to write the screenplay with her partner, the writer-director Noah Baumbach,” Barasch writes. “Mattel and Warner Bros. insisted on seeing a preview of the script’s contents, but the couple balked—they needed the freedom to experiment. Jeremy Barber, an agent at U.T.A. who represents Gerwig and Baumbach, is close with Brenner, so he could be blunt. ‘Are you crazy?’ he told her. ‘You should’ve come into this office and thanked me when Greta and Noah showed up to write a f--king Barbie movie!’”