Can leaders use GenAI too much? The hidden risks of AI-driven leadership and decision-making

Can over reliance on Gen AI undermine leadership capability?

Generative AI has rapidly moved from an experimental technology to an executive operating tool. Senior leaders now use GenAI systems to draft reports, summarize meetings, analyze markets, prepare speeches, review policy options, generate strategic plans, and support investment decisions. In many organizations, AI has become embedded into the daily workflow of leadership itself.

The productivity gains are real. Tasks that once took hours can now be completed in minutes. Managers can process larger amounts of information, generate multiple strategic scenarios, and communicate more quickly across increasingly complex organizations. Governments are using GenAI to support policy analysis and service design. Businesses are deploying it to accelerate product development, improve forecasting, and reduce administrative overhead.

Yet an important question is emerging beneath the excitement: can leaders use GenAI too much?

This question matters because leadership is not simply an information-processing exercise.

Effective leadership depends on judgment, emotional intelligence, skepticism, strategic interpretation, ethics, institutional knowledge, political awareness, and human trust. If leaders become overly dependent on AI-generated analysis and communication, there is a growing risk that some of the very capabilities that define strong leadership could weaken over time.

There are already signs that this may be happening. Research increasingly suggests that excessive reliance on GenAI can reduce critical thinking, create overconfidence, weaken collaboration, erode trust between managers and staff, and distort decision-making processes. In some cases, AI systems may amplify optimism bias or encourage leaders to rely too heavily on probabilistic answers that sound convincing but lack real strategic depth.

This creates a difficult balancing act for modern organizations. Leaders who ignore AI risk falling behind competitors and peers. But leaders who rely on AI too heavily may unintentionally weaken organizational resilience, creativity, and institutional capability.

The challenge is no longer whether organizations should use GenAI. The real issue is how leaders can use it without becoming dependent on it.

Why executive over-reliance on GenAI is becoming a strategic risk

The rapid adoption of GenAI among executives is creating a major shift in organizational power structures. Senior leaders generally have the greatest access to AI tools, the strongest digital literacy support, and the most freedom to experiment with new technologies. Meanwhile, frontline staff and individual contributors often lag behind.

This imbalance matters because it can create a two-speed organization. Strategy becomes increasingly AI-enabled at the top while operational execution remains fragmented below. Executives may assume transformation is progressing rapidly because their own productivity has improved, even while large sections of the workforce struggle to integrate AI into daily operations.

Research increasingly shows a significant adoption gap between executives and frontline workers. Managers and executives report much higher levels of regular GenAI use compared to individual contributors. At the same time, many employees do not fully understand how AI tools are selected, governed, or monitored within their organizations.

This creates several risks simultaneously.

First, leadership teams may develop unrealistic assumptions about workforce readiness. Second, employees may begin to see AI as a top-down control mechanism rather than a collaborative productivity tool. Third, uneven access to AI capabilities may deepen existing organizational hierarchies instead of democratizing innovation.

The issue becomes even more complex in large public-sector organizations and multinational enterprises where transformation speeds vary across departments, sectors, and regions.

Financial services firms may move rapidly into AI-supported operations while healthcare, retail, education, or government agencies adopt AI more cautiously due to regulatory, ethical, or operational constraints.

Leaders therefore face a paradox. AI can improve coordination and decision-making speed, but if adopted unevenly it can also increase fragmentation inside organizations.

The danger of AI-assisted overconfidence

One of the most important emerging concerns is the relationship between GenAI and executive overconfidence.

Recent executive experiments involving forecasting exercises provide important insights into how AI may influence managerial thinking. In one study involving hundreds of managers and executives, participants were asked to forecast the future stock price of a well known company after reviewing market trends. One group discussed their forecasts with peers, while another group consulted ChatGPT before revising their predictions.

The results were revealing.

Executives who used ChatGPT became significantly more optimistic in their forecasts. Those who discussed with peers became more cautious and conservative. More importantly, the AI-assisted group ultimately produced less accurate predictions compared to both their original estimates and the peer-discussion group.

This finding matters far beyond stock forecasting. It suggests that GenAI may unintentionally encourage leaders to become more confident in uncertain environments while simultaneously reducing forecast accuracy.

Several psychological dynamics appear to drive this effect.

AI authority bias

GenAI systems often communicate with extraordinary fluency and confidence. Their responses appear detailed, analytical, and comprehensive. For many executives, this creates a form of “AI authority bias” where the sophistication of the language increases perceived credibility.

Leaders may unconsciously give more weight to AI-generated analysis simply because it sounds polished and data-driven.

This is particularly dangerous in complex strategic environments where confidence and accuracy are not the same thing. A persuasive AI-generated recommendation can appear highly rational while still being incomplete, biased, or fundamentally flawed.

In traditional organizations, uncertainty is often moderated through debate, disagreement, and peer challenge. AI systems do not naturally replicate these social dynamics unless leaders deliberately structure them into the decision-making process.

Trend extrapolation and false certainty

GenAI models are fundamentally pattern-recognition systems trained on historical information. They are often highly effective at identifying trends and generating probabilistic outputs. However, this can create problems in environments where disruption, shocks, political events, or unexpected behavioral changes matter more than historical continuity.

In the forecasting experiment, AI systems appeared to reinforce recent upward momentum rather than challenge underlying assumptions.

This is a critical insight for strategic leadership. Many of the most important business and policy decisions involve turning points rather than trend continuation. Economic crises, geopolitical instability, regulatory intervention, technological disruption, social backlash, and market corrections often emerge precisely when historical patterns break down.

Human intuition, skepticism, and emotional caution sometimes provide an important counterweight to purely data-driven reasoning. AI systems, by contrast, may continue extending existing patterns further into the future than is strategically justified.

The illusion of knowledge

Another growing concern is the “illusion of knowledge” created by GenAI systems.

Because AI can instantly produce large volumes of coherent information, leaders may begin to feel more informed than they actually are. Access to sophisticated language and rapid synthesis can create the impression of deep understanding even when critical contextual knowledge is missing.

This creates a dangerous environment for executive decision-making.

Leaders may spend less time interrogating assumptions, seeking dissenting opinions, or testing alternative scenarios because AI-generated outputs create a false sense of analytical completeness. Over time, this can reduce organizational curiosity and weaken strategic challenge functions inside leadership teams.

How excessive AI use can weaken critical thinking

One of the least discussed risks of GenAI adoption is cognitive atrophy.

When leaders rely too heavily on AI-generated summaries, recommendations, and communications, they may gradually reduce their own engagement with deep analysis and independent reasoning.

This is not simply a theoretical concern. Human cognitive skills develop through repetition, practice, and active problem-solving. If AI increasingly performs the tasks associated with synthesis, drafting, analysis, and interpretation, leaders may begin outsourcing parts of their thinking process itself.

This can manifest in several ways.

Leaders may stop reading source materials in depth because AI summaries appear sufficient. Managers may rely on AI-generated presentations rather than conducting original analysis. Strategic planning sessions may become increasingly shaped by AI-generated frameworks instead of internal debate and reflection.

Over time, organizations risk producing leaders who are highly efficient but less intellectually rigorous.

The danger becomes especially pronounced in governments and large institutions where institutional memory and nuanced judgment matter deeply. Public policy decisions often involve ethical trade-offs, competing stakeholder interests, political realities, and long-term societal consequences that cannot be reduced to pattern recognition alone.

Similarly, in corporate environments, leadership frequently requires navigating ambiguity rather than optimizing clear variables. Human judgment remains essential when dealing with culture, trust, negotiation, crisis management, geopolitical risk, and public perception.

AI can support these functions, but it cannot replace the deeper human reasoning processes required to lead through uncertainty.

The erosion of trust and authentic leadership

Leadership is fundamentally relational. Employees do not simply evaluate whether managers are productive; they evaluate whether leaders are authentic, empathetic, trustworthy, and emotionally engaged.

This creates another major risk associated with excessive GenAI use.

When employees believe managers are outsourcing communication, feedback, or recognition to AI systems, leadership can begin to feel synthetic rather than human.

Many organizations are already experimenting with AI-generated performance reviews, automated employee feedback systems, AI-assisted recognition messages, and AI-generated internal communications. While these tools may improve efficiency, they also risk weakening the emotional credibility of leadership interactions.

Employees are often highly sensitive to authenticity. A performance review that appears polished but emotionally generic may undermine trust more than a shorter but genuinely personal conversation.

This issue becomes even more significant among younger workers. Many Gen Z employees are enthusiastic users of AI technologies while simultaneously expressing skepticism about how organizations deploy them. Younger professionals increasingly expect transparency, ethical governance, and authentic communication from employers.

If leaders appear to hide behind AI-generated messaging, trust deficits may deepen further.

The long-term risk is cultural fragmentation. Organizations may become operationally faster while simultaneously becoming emotionally weaker.

Pressure on managers and burnout risk

Managers are emerging as one of the most pressured groups in the GenAI transition.

Senior executives often focus on enterprise-level transformation, while frontline employees focus on adapting to new workflows. Managers sit uncomfortably between these two worlds.

They are expected to:

• learn new AI systems,

• maintain operational performance,

• guide team adoption,

• manage employee anxiety,

• enforce governance standards,

• redesign workflows,

• and continue delivering results.

Research shows that a large majority of managers report significant workload changes due to AI adoption. Many are also expected to acquire entirely new technical and leadership capabilities at high speed.

This creates what some analysts now call the “manager squeeze.”

Managers become both transformation targets and transformation enablers simultaneously.

The pressure is particularly acute in government agencies and large enterprises where digital transformation programs already compete with budget pressures, regulatory obligations, staffing shortages, and political scrutiny.

Without proper support, the result may be managerial burnout disguised as digital progress.

Organizations often assume managers are coping because operational performance remains stable in the short term. In reality, many managers may simply be absorbing additional stress privately while projecting confidence publicly.

This has important strategic implications. If middle management becomes overwhelmed or disengaged, GenAI transformation programs may stall regardless of executive ambition.

Operational and governance risks of excessive AI dependence

Beyond leadership quality and organizational culture, overreliance on GenAI creates direct operational and governance risks.

Bias and discrimination

AI systems inherit biases from training data and historical patterns. If leaders rely excessively on AI-generated recommendations in hiring, promotion, performance evaluation, or workforce management, organizations may unintentionally reinforce existing inequalities.

This is particularly concerning in public-sector environments where fairness, accountability, and procedural transparency are essential.

Even well-intentioned leaders may struggle to explain or defend AI-supported decisions if underlying models lack transparency.

Confidentiality and security risks

Many leaders underestimate the security implications of using public AI tools.

Sensitive corporate, government, legal, financial, or personnel information may be inadvertently exposed when entered into external AI systems without appropriate safeguards.

As AI adoption accelerates, organizations increasingly need formal governance frameworks covering:

• approved AI platforms,

• data handling protocols,

• confidentiality requirements,

• audit mechanisms,

• and escalation procedures.

Without clear guardrails, the convenience of AI can create major compliance and reputational risks.

Strategic homogenization

Another emerging risk is strategic sameness.

Because many organizations use similar AI tools trained on broadly similar datasets, there is a growing possibility that strategic thinking itself becomes more standardized.

If leadership teams increasingly rely on AI-generated frameworks, recommendations, and market analysis, competitive differentiation may weaken over time.

This matters because long-term competitive advantage often comes from unconventional thinking, institutional culture, and original strategic insight rather than optimized averages.

AI systems are generally designed to predict likely answers, not generate truly disruptive thinking.

Organizations that become too dependent on AI-assisted consensus may unintentionally reduce their own strategic creativity.

Why peer discussion still matters in the AI era

One of the most striking lessons from recent executive experiments is that peer discussion often improved judgment quality more effectively than AI consultation alone.

Human discussion introduces:

• skepticism,

• emotional caution,

• contextual awareness,

• dissent,

• ethical questioning,

• and social calibration.

These dynamics can prevent organizations from moving too quickly toward overly optimistic or narrowly technical conclusions.

This does not mean human groups are always superior. Human decision-making can also suffer from groupthink, politics, ego, and bias.

However, the evidence increasingly suggests that the best outcomes emerge when AI augments human dialogue rather than replacing it.

Organizations should therefore avoid treating AI as a substitute for collaborative leadership processes.

The most effective leadership models will likely combine:

• AI-assisted analysis,

• structured human debate,

• diverse stakeholder input,

• and independent challenge functions.

This hybrid approach allows organizations to capture the efficiency of AI without losing the corrective value of human interaction.

Are concerns about GenAI overuse are overstated?

It is also important to consider the opposite argument.

Some analysts argue that concerns about AI overreliance may be exaggerated and reflect a broader historical pattern of resistance to new technologies.

Throughout history, leaders have worried that calculators would weaken mathematics skills, search engines would reduce memory, and computers would diminish analytical capability. Yet productivity and knowledge access ultimately expanded dramatically.

From this perspective, GenAI may simply represent the next stage of cognitive augmentation.

Supporters of aggressive AI adoption argue that:

• leaders should delegate routine thinking to machines,

• productivity gains free humans for higher-order work,

• AI can reduce decision bottlenecks,

• and organizations that hesitate may lose competitiveness.

They also argue that AI systems are likely to improve rapidly. Current limitations around hallucinations, bias, and contextual awareness may become less severe as models become more advanced, connected to real-time data, and integrated into enterprise governance systems.

In this view, the greater risk may actually be underutilization rather than overutilization.

Organizations that avoid deep AI integration could become slower, less adaptive, and less competitive than AI-enabled rivals.

There is some truth in this argument. AI undoubtedly provides substantial value in many contexts. Leaders who completely reject GenAI may struggle to keep pace with the speed and complexity of modern environments.

However, even advocates of aggressive adoption increasingly acknowledge that governance, human oversight, and critical thinking remain essential.

The debate is therefore not really about whether AI should be used. The real issue is how organizations preserve human judgment while benefiting from machine intelligence.

Building a sustainable hybrid leadership model

The future likely belongs neither to fully AI-driven leadership nor purely traditional leadership models.

Instead, organizations will need to build forms of “hybrid intelligence” where AI enhances rather than replaces human capability.

Several principles appear increasingly important.

Treat AI as a co-pilot, not an autopilot

Leaders should view AI outputs as starting points for analysis rather than final answers.

AI-generated recommendations should be challenged, debated, stress-tested, and contextualized before decisions are made.

Protect human interaction

Organizations should preserve spaces for genuine conversation, mentoring, and collaborative thinking.

Not every communication process should be automated simply because it can be.

Invest in critical thinking

As AI capabilities expand, human critical thinking may become more valuable rather than less valuable.

Organizations should actively train leaders and employees to:

• question AI outputs,

• identify bias,

• test assumptions,

• and recognize uncertainty.

Support managers properly

Manager enablement should become a core component of AI transformation programs.

This includes:

• realistic workload expectations,

• peer learning networks,

• leadership training,

• governance guidance,

• and mental health support.

Build transparent AI governance

Trust will become a major competitive advantage in the AI era.

Organizations that clearly explain:

• how AI is used,

• how decisions are made,

• what safeguards exist,

• and how employees are protected

are likely to build stronger long-term workforce confidence.

Leadership still requires human judgment

Generative AI is already reshaping leadership, management, and organizational decision-making. Its ability to accelerate analysis, communication, and productivity is undeniable. Used well, GenAI can help leaders process complexity faster and support better operational performance.

But there is a growing danger in assuming that faster thinking automatically means better thinking.

The evidence increasingly suggests that excessive reliance on AI can contribute to overconfidence, weaker critical thinking, reduced collaboration, skill atrophy, and declining trust between leaders and employees. AI systems can produce convincing answers without truly understanding context, ethics, politics, or human emotion.

Leadership is ultimately more than prediction and efficiency. It involves judgment under uncertainty, emotional credibility, institutional memory, and the ability to navigate competing human interests.

The organizations most likely to succeed in the GenAI era will not be those that replace human leadership with AI systems. They will be the ones that combine machine capability with human wisdom more effectively than their competitors.

The challenge for modern leaders is therefore not deciding whether to use AI. The challenge is ensuring they do not lose the distinctly human capabilities that make leadership valuable in the first place.

For more insights on digital transformation, AI strategy, public-sector innovation, and emerging global trends, subscribe to additional articles and analysis at George James Consulting