For decades, boards have learned to live with disruptive technologies—ERP, CRM, cybersecurity, the cloud. AI feels different, says veteran director, tech CEO and former IBM executive Cheemin Bo‑Linn. “It touches every function and every decision the board is supposed to oversee.” In her view, AI is no longer just another system to be procured; it is a pervasive capability that boards must understand if they are to meet their obligations.
“If you don’t have a basic understanding of AI and digital technologies, how can you perform your duty of care and duty of loyalty?” Bo‑Linn asks. As a former CEO and Fortune 100 tech executive who now serves as a board chair and lead independent director, she sees AI moving quickly into the standard of care against which directors will be judged. Boards are using it to compress 100‑plus‑page board books into decision‑ready intelligence and to surface anomalies management might gloss over. At the same time, she notes, “Governments are putting laws and frameworks in place on the misuse of AI,” from bias in lending and housing decisions to deceptive consumer practices—raising the stakes for boards that remain passive.
Bo‑Linn is clear that directors do not need to become technologists. “Do you need to be a cybersecurity expert to oversee cyber? No. It’s the same with AI,” she says. “You don’t need to code, but you do need enough foundational understanding to see the high‑level implications for the company.” That means knowing where AI could materially improve performance, where it could create new vulnerabilities and how it changes the economics of strategy and risk.
She likens this moment to the early days of cyber on the board agenda. “With cybersecurity, the question was never ‘if,’ it was ‘when’ you would be attacked,” she says. “With AI, it’s ‘when’ it will disrupt your industry, not ‘if.’” The boards that still treat AI as a side project or a niche IT issue, she warns, are repeating the mistakes of the cyber laggards. “Some people still have a traditional way of thinking, but this is so pervasive, so encompassing, you have to have clean‑slate thinking to understand it.”
1. Turn board books into intelligence.
One of the most tangible shifts Bo‑Linn sees is in how boards consume information. “We used to have a 107‑page board packet in the portal,” she recalls. “You’d hope everyone read it, but to really synthesize it was hard. Each line executive would present their piece, and it just went on.” AI is beginning to change that equation. “Now, AI can condense those materials and bubble up the key issues, opportunities and risks,” she says. “It takes data and turns it into information—and then into intelligence for the board.”
Crucially, she believes AI helps directors see what isn’t being said. “When a C‑Suite executive presents, they usually focus on the wins and maybe the bittersweet parts,” she says. “AI will crawl through everything—Salesforce, yield data, win–loss reports—whether you want it to or not, and say, ‘This is what’s really happening.’” That can mean surfacing a major customer that has deferred orders for six straight quarters or revealing that apparent yield gains are simply seasonal effects. “Numbers don’t lie,” she says. “They can be skewed, but they don’t lie. AI makes it much harder for anyone to hide.”
2. Ask for options B and C.
If AI is one part of the equation, mindset is the other. Bo‑Linn sees a spectrum of board behavior: “Some are hesitant because they don’t have enough education; they’re so consumed with the risk that inaction becomes the decision,” she says. “On the other side, some are overly optimistic. This is not a magic wand—it has some heroes, but it has more fallen heroes.” Her advice: “Fail better and fail forward. You will make mistakes. Make sure they’re recoverable.”
For boards that want to move beyond hype, she recommends a simple discipline: always asking for alternatives. “When management comes in with Strategy A, the board should be asking, ‘What were B and C? What did the data tell you about those?’” AI can help management model different scenarios and compare options under varying assumptions. “If you don’t ask those questions, you’re not getting the full benefit of AI,” she argues. “You’re just decorating the existing plan instead of challenging it.”
3. Learn from auditors—without outsourcing judgment.
Many directors are first encountering AI through their auditors. “A lot of the big firms are already using automation and AI to find anomalies and do smarter sampling,” Bo‑Linn notes. “That’s good—they promote it as part of their differentiation.” But she cautions boards not to confuse that with a comprehensive AI strategy. “Audit partners look at the past and the present. They’re not responsible for the future,” she says. “You can’t say, ‘Our auditors are using AI, so we’re covered.’”
Consultants and advisors, she points out, “come in and they go out. They are not there to continually monitor.” Boards, by contrast, have recurring meetings and standing committees that see patterns over time. “At least one person on the board needs enough experience to ask the right questions and translate what it all means, good or bad,” she says. “Otherwise, the timing is not in your favor—competition is moving, and you’re standing still.”
4. Be aware of the legal and ethical stakes.
As AI becomes more embedded in core processes, the legal and ethical stakes rise. “My concern is: Who’s watching the kids?” Bo‑Linn says with a laugh. She points to regulators in the U.S. and Europe that are scrutinizing how AI is used in lending, housing and other consumer decisions. “AI can rate you based on where you live, your gender, your age. If you don’t understand how those models work, you could end up with discrimination built into your system.”
That’s not a risk boards can delegate away, in her view. “Board members cannot passively delegate their decision‑making on AI,” she says. “We only have three major responsibilities: deliver shareholder value, hire and fire the CEO, and approve strategy and risk oversight. AI now cuts across all three.” If directors do not take those responsibilities seriously, she warns, “they are going to be open to a lawsuit. And some aspects of it may not be fully covered by D&O insurance.”
Data quality is another flash point. “If you don’t have good data hygiene, the AI is not going to work,” she says. “AI will unveil your infrastructure weaknesses.” That’s why she argues for strong data governance and clarity about how third‑party tools use company data. “You need rules for how people use large language models—what data is acceptable, which tools are allowed,” she says. “Data rules must be in place for AI just like any other tool.”
What Directors Can Do Now: A Practical To-Do List
Given those stakes, what should directors be doing now? First, Bo‑Linn says, boards should look at their own composition. “You don’t need or want a pure AI engineer or researcher on the board—that’s too limiting,” she says. “Look for someone who comes from tech, has run businesses and understands other industries. A former GM or operator with AI and cyber knowledge is far more valuable in the boardroom.” Narrow roles like CSO or CIO, she adds, may not have the breadth of experience boards need across all topics they oversee.
Second, she advocates for raising the baseline of literacy around the table. “Everyone should be using AI tools daily in their lives to understand what they can and cannot do,” she says. Formal education matters too: “You have to have enough foundational understanding of digital technologies like AI and generative AI to at least understand the risks and, more importantly, the opportunities.”
Finally, Bo‑Linn urges boards to treat AI as an enterprise‑wide transformation, not a series of pilots. When she was CEO, she says, “I had to reimagine how the organization, the work plan, product development, manufacturing, supply chain—eight or 10 different functions—would look in an AI world.” That kind of reimagining takes time and capital, and boards need to demand clarity on both. “My board would ask, ‘You want how much—for what? And what does success look like?’” she recalls. “Any CEO who says they started an AI project in January and have captured all the value by December is fooling themselves. The real value is enterprise‑wide, sustainable and repeatable—and that’s exactly why the board has to be engaged.”
