Leadership has always been a lonely endeavor, but in the AI era, that loneliness is at risk of getting worse—and when it does, the negative impact on CEOs and leaders jeopardizes the future of their organization.
Leadership isolation degrades the product of leadership itself: the quality of decisions, the coherence of strategy, the ability to execute at scale. It takes much more than a wellness initiative to address it, requiring a careful strategic approach from the top—for the betterment of individual leaders, and the organizations, products and people depending on them.
More than half of CEOs report experiencing isolation, and most say it affects their performance. Loneliness is often misinterpreted as a personal issue, such as a need for more connection in the casual sense, like knowing your teammates’ children’s names, or what they did on the weekend. In reality, loneliness is not about being alone, but about thinking alone. It’s a breakdown in psychological safety where leaders lack the candid challenge needed to refine decisions as they are made, a finding reinforced at scale by Google.
Erosion of Psychological Safety
At the same time, emerging research shows that increased interaction with AI is associated with greater loneliness and reduced social connection, and early evidence suggests AI adoption may further erode psychological safety in organizations.
The risk isn’t the advancement of AI in itself, but that it reinforces isolated thinking at the very moment leaders most need diversity of thought, dissent, debate, collaboration, correction and the ability to align to achieve strategic goals.
Under pressure, leaders can unintentionally make problem-solving feel dangerous when psychological safety is not actively demonstrated. Signals such as tone, impatience or absence lead teams to withhold honest guidance and pushback, reducing the flow of critical thinking and candid input. This exacerbates leadership isolation from the CEO throughout the organization, resulting in decisions formed with limited scrutiny and a widening gap on vision, ownership and success. Without psychological safety, this pattern is difficult to reverse and directly impacts outcomes and the overall success of the organization.
Unfortunately, most organizations have ultimately designed the conditions that have produced this very situation. Engagement-driven systems optimize low-friction interaction, flattening human dynamics that contribute to the production of isolation itself. Over time, reduced feedback narrows perspectives, reinforces existing assumptions and distances leaders from the realities of their organization. Too often, I hear CEOs ask, “Why aren’t they sharing this information with me?” when the more important question is “What in the system is making it unsafe to do so?”
AI is accelerating this cycle. As AI is embedded into a leader’s immediate thinking and judgement, it introduces a feedback system that is fast, frictionless and independent of human challenge.
We are now facing a dual challenge: Alongside AI use contributing to increased feelings of loneliness, a broader loneliness epidemic has been identified at a societal level.
When AI rapidly changes the conditions under which leadership thinking occurs, we experience a paradox—the tool that accelerates thinking also reduces exposure to the human challenge required to test, refine, evolve and execute it.
Leaders are naturally drawn to rapid ideation and problem-solving through AI, yet more of that process happens in isolation. What was once developed through dialogue is increasingly shaped through interaction with AI systems that do not test assumptions or introduce necessary friction from the team.
A Widening Gap
Each major shift in communication—from the telephone to email to mobile—has required leaders to adapt to a reduction in social cues that once shaped how the meaning of a challenge or opportunity was interpreted and shared.
Research on email communication showed that even when people wrote with strong emotional intent, recipients filtered messages through their own mental models, experiencing them as ambiguous and requiring further clarification. Humans consistently overestimate how well their ideas land.
In today’s volatile economic, environmental and geopolitical environment, this creates a new leadership challenge that AI does not currently resolve.
Instead, by introducing a cognitive partner that allows leaders to develop thinking in isolation, AI increases the gap between how leaders believe their thinking is understood and how it is actually received. This widens the disconnect between intention and alignment, reducing opportunities for ideas to be challenged, co-created and refined for execution at scale.
Over time, this deepens systemic isolation between leaders, employees, across teams and throughout the organization.
What Leaders Are Living Through
To understand how this dynamic plays out in practice, I spoke with three senior leaders navigating AI adoption inside fast-moving organizations and a veteran management consultant who have seen failure and what works.
One pattern recurred: A leader or CEO has a one-off, impressive experience with an AI tool and immediately generalizes it into an organization-wide mandate. “They tried it once and thought they understood it universally,” said one C-level leader for a major technology brand in London, who asked to remain anonymous. “That’s where the isolation compounds. They’re not listening anymore—they’re dictating and forcing others to isolate, compete then burnout. The cycle is right before our eyes.”
This C-level leader described a situation in which a chief creative officer under pressure from a CEO instructed the entire division—designers, copywriters, creative directors—to build a minimum of two agents within a month, with no definition of what problem or mission those agents were meant to solve. The directive cascaded from the top; no one asked what the actual friction was.
The mandate brought confusion, disengagement, conflict and attrition among the people best positioned to use AI well. The employees who challenged it were ignored and told to get on with it; those who complied but didn’t understand why burned out. Today we see many examples of organizations where AI usage dashboards surveil activity instead of assessing contributions to the mission, the problem or the bottom line.
The dynamics illustrate how AI adoption, when implemented without a clear understanding of what business problem needs to be solved, deepens the isolation it was meant to resolve. A 2025 peer-reviewed study published in Humanities and Social Sciences Communications found that AI adoption significantly reduces psychological safety, which in turn increases rates of depression among employees, hampering creativity, collaboration and problem solving—a declining pathway that intensifies when ethical and emotionally intelligent leadership is absent. The layered consequence: Leaders who push AI mandates without building psychological safety are simultaneously creating catastrophic conditions that make those mandates fail.
A Better Way
James Pycock, VP of Product at Albert, a Series A AI company in San Francisco, offered a contrasting perspective. Everyone in his organization is a builder—from engineers, designers and product managers now classified under a single “build organization.” “There’s a kind of reassurance in it,” he said. “Yes, our roles are going to change and evolve. But what we all do—what holds us together —is what we ‘build.’”
Pycock observed something counterintuitive to our greatest concerns about AI in organizations: As AI takes over production work, leaders are becoming more, not less, human in their day-to-day presence. “I can imagine myself spending more time…one on one with people who report to me,” he said. “Because production is faster and easier now. The thinking work, the relational work—that’s what’s left. And the most important thing for us to solve are our customers problems.” His company has also introduced a chief work officer role—a function whose mandate is to bring AI-driven efficiency across the organization using a “Formula One pit crew” model: The head of each function still drives but now has a support crew rotating in to accelerate execution.
This structural reinvention points to a broader behavioral opportunity for organizations to stop treating AI as a solo productivity tool and start designing it as an infrastructure for collective work, reversing the isolation dynamic. Pycock’s teams now return to first principles when interviewing, hiring for grit, product taste and judgement, not for knowledge that AI can generate. “Why are you asking questions that ChatGPT can answer?” he noted. “Interview for things it can’t. We now interview people in person to see that in action.”
Play Like a Crow
Melissa Swift, author of Effective and a veteran of Korn Ferry, Mercer and Cap Gemini, offered a behavioral lens reframing what organizations typically misread as resistance. “We’ve positioned AI as an anti-social experience,” she observed. “And then we call it change resistance. But it isn’t. Employees are just not enjoying themselves.”
Swift points to research on behavioral research on crows—animals that, when given the option, actively prefer to use tools to complete tasks, even when they don’t have to. The implication is that technology use, far from being inherently alienating, is naturally pro-social and rewarding, but only when it involves some degree of agency, play and shared purpose. “Think about why people love video games,” she said. “They’re playing with others. It’s pro-social. We’ve made AI into the opposite of that.”
Swift describes one organization where a training program did the opposite of what it intended. Employees were required to complete AI training under threat of pay docking, but then denied access to the actual tools. On some client engagements, staff were expected to simulate AI generated outputs they weren’t licensed to produce. The training had become compliance theater, surveillance in place of support and performative adoption instead of skill building that would have built trust and reduce the curve of burnout and isolation.
The CEO’s Next Steps
AI is changing leadership thinking almost daily. To operate effectively in this evolving new environment, leaders must redesign how thinking happens across the organization.
This requires a systems-led approach led by the CEO where AI and human integration can thrive—demonstrated with candor and anchored by reimagined leadership behavior fit for purpose in the AI era.
This requires structured and synchronized reinvention:
Mea Culpa—Reset Leadership Assumptions, Expectations and Mindset
Start with candid acknowledgment that what got you here will not get you there. Create space for honest reflection with your leadership team on how communication and decision-making have broken down and where leadership behavior has contributed to current norms. This requires leaders to confront how their inherited behaviors, systems and beliefs have shaped current operating conditions. It also requires leaders to create fundamentally different conditions. This is where psychological safety begins—enabling the team to re-engage to redefine ownership and intention for this era. A trusted executive coach can help reinvent your C-Suite’s operating system.
Tabula Rasa or Blank Slate—Redesign Communication Architecture
Start over. Rebuild how ideas are challenged, tested and refined. Design systems where disagreement and first principles are expected, not avoided, and where challenge is positioned as a way to better the end result versus a risk to be managed.
AI Integration—Balance AI Intelligence Speed with Healthy Human Intelligence Friction
James Pycock’s model at Albert offers a practical template: Use AI as a tool for acceleration but ground it with human intelligence where thoughts can be shared and outcomes debated. Set guardrails to avoid early symptoms of burnout and protect the team’s cognitive and mental integrity for optimal decision-making. Ensure that critical decisions pass through real-time dialogue, feedback and relational presence rather than being formed solely through individual cognition.
Design for Play (Not Compliance)
The organizations seeing the most sustainable AI adoption create conditions for curiosity. Rudi Angonno, formerly of Google and LEGO, recounts being one of the leaders who started culture change from the ground up using volunteer-driven facilitation and bottom-up principle definition rather than top-down edict—an approach MIT Sloan Review documented as a driver of genuine leadership culture change. The practical implication: Before deploying AI at scale, start with a low stakes pilot in a domain that matters to someone specific. Find the problem and let people discover what the tool can do for their work, not the quarterly targets.
At LEGO, Angonno worked with a studio artist whose day job was narrow and at risk of automation. Rather than issuing an AI mandate, he identified a real creative need—original music for a children’s product line—and recognized that the artist, also a musician, could use AI audio tools to generate scored tracks that no one else on the team could produce. “I basically converted him,” he recalled. “Now you can bring your whole self into the work. You’re still a studio artist. But…guess what: you just added one more skill set to the company that will benefit in numerous ways.” The artist moved from a narrowly defined, easily automated role to a cross-functional creative specialist with a genuine competitive advantage. The project succeeded by starting small, low-risk and focused on a real problem rather than a mandate.
As author Melissa Swift frames it: “Play might actually be the fastest path to ROI, not the slowest.”
Catalyst-Citizen Model—Distribute Ownership, Ideation and Responsibility
Psychological safety cannot rest solely with the leader. It must be embedded and integrated horizontally and vertically, with individuals acting as challengers (Catalysts) and stabilizers (Citizens), enabling both creative tension and cohesion to reach organizational outcomes. In practice, this looks like Albert’s “everyone is a builder” reframe—collapsing boundaries and creating a new shared identity between engineering, product and design, making collaboration and focus natural. Similarly, NVIDIA’s “Mission is the Boss” system removes silos associated with traditional companies—a principle that becomes more powerful as AI spreads throughout an organization.
100-Day Reinvention Sprint—Operationalize Change
Organizational adaptation at NAVI or BANI speed cannot be sustained by leaders operating under chronic stress. Cognitive clarity and relational presence are physiological capacities requiring a structured 100-day sprint focused on how the organization behaves collectively to create the conditions for sustainable change alongside AI.
A technological environment optimized for efficiency alone catastrophically reduces the human challenge on which leadership depends.
Loneliness at the top was already a performance risk before AI arrived. Now AI is the accelerant. It’s more pervasive than social media, a thought partner in everyone’s pocket and embedded in every organizational dashboard.
Leaders who adopt AI most aggressively, or who think the antidote to leadership isolation is a culture survey hiding behind numbers, will not be the ones who will reinvent their destiny. Instead, those who succeed will recognize that the antidote is a deliberate reinvention of how their organization thinks and collaborates symbiotically human-to-human and human-to-AI without eroding psychological, emotional and relational integrity, driving positive outcomes for shareholders, customers and employees alike.
The CEOs beating isolation will do it by building together with AI, playing together with AI and thinking together with AI—changing everything about how their organizations perform.
This raises the question worth putting to every board: If collaboration became a primary organizational KPI—measured, reported and weighted alongside efficiency and output—would loneliness still accelerate the way it does? Organizations that treat connection as a metric, not a mood, may very well see the results in this age of AI, climate and societal disruption.
