From board decks to earnings calls to leadership offsites and coffee-machine conversations, the topic of AI is ubiquitous. The opportunity is enormous: to reimagine work, unlock creativity, and expand what organizations and people can do. So is the pressure.
In response, many organizations are rolling out tools and launching pilots. Some of this activity is necessary. Much of it, however, misses the deeper point. Too many leaders are asking: how will AI change us? The better question is: what kind of leadership will we build to guide AI?
That distinction matters because technology alone does not shape outcomes. Leadership decisions do—meaning the systems, norms, and capabilities that organizations choose to build and apply to their work.
Here are three ways to strengthen what people can bring to the table in the age of AI.
Don’t allow fear to shrink ambition
AI’s promise lies in bold experimentation. Even in the most sophisticated organizations, however, fear is quietly constraining it. So there is tension. Leaders ask their people to make intrepid experiments with AI, while launching efficiency programs that employees interpret as precursors to job cuts. When people feel exposed, they play small. Breakthrough ideas give way to micro use cases and firms refine today’s’ model instead of creating tomorrow’s.
What to do: Leaders can reduce fear by creating a protected space for AI experimentation, shielded from short-term efficiency pressure. Research has found that such psychological safety is critical to performance. Teams that feel secure identify problems earlier, challenge assumptions more freely, and learn faster. If leaders want bold thinking, they must lower the perceived cost of offering it. Otherwise, AI may improve efficiency while the reimagining moment slips by.
History proves the point. When Siemens and Toyota were reinventing their production systems, they explicitly protected jobs. What the companies gave up in short-term savings, they gained in long-term innovation. People were emboldened to take risks because they believed productivity benefits would be shared, not weaponized.
Creating opportunities for people to learn is another way to help to reduce fear and liberate people to think beyond the readily possible. That was the thinking behind CEO Satya Nadella’s effort to instill a “learn it all” mindset at Microsoft; this made it okay to not already know it all and contributed to breakthroughs in product and strategy. Another approach is to offer regular time for generative work, such as Google’s “20% time” practice, in which engineers were encouraged to explore personal projects that could help the company. AdSense and Google News, among other products, began this way.
Use AI as an input, not a default
From the wheel to yesterday’s AI agent, every invention has either augmented or replaced human actions. The danger is when people rely on the tool so much that they stop thinking.
As access to AI models and computational power spread, analytical advantages erode. That makes the distinctive human ability to interpret context, weigh trade-offs, understand stakeholder impacts, and question outputs even more valuable. Stanford’s Human-Centered Artificial Intelligence institute has found that teams combining AI recommendations with expert oversight consistently outperform fully automated systems. Or, as my son’s first-grade teacher put it: being smart is knowing a tomato is a fruit. Being wise is knowing not to put a tomato in a fruit salad.
What to do: Design decision-making to ensure that AI informs judgment rather than replaces it. For major decisions, leaders should require teams to document the human reasoning behind AI-informed decisions, making the logic explicit so that it can be tested. Over time, this builds discernment and institutional memory, and ensures that people take responsibility for their calls, rather than blaming the models. Teams can also foster structured dissent as a counterweight to AI-driven overconfidence by asking questions like, “What would have to be true for this to hold?”
Keep humans at the center of value judgments
Ethical leadership in the AI era is about deciding, explicitly and repeatedly, where optimization must stop and human responsibility must begin. Among the questions to be considered: What decisions should algorithms be allowed to make? Who is accountable when an AI-based decision causes harm?
What to do: It’s important for leaders to articulate what lines will never be crossed. Embed governance into workflows, ensuring people make the most important decisions; train managers to weigh what is possible against what is responsible.
Judgment, ethics and values cannot be outsourced to AI. These capabilities must be built, then tended, so that they become second nature—starting from the top but imbedded throughout the organization. In business, trade-offs are inevitable; in the age of AI, they need to be intentional.
The leaders who get this moment right will not deploy AI tools just because they can; they will do so in a way that tap into psychological safety, human judgment, and ethical clarity. Efficiency without empathy is not progress. Innovation without judgment is not leadership.
AI won’t decide the future. Leaders will—and history will be unforgiving about the difference.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
