Leadership development has not caught up with AI. The programmes, frameworks, and competency models that organisations use to develop leaders were designed for a pre-AI world. Gerson's research across 40 organisations reveals a consistent gap: leaders know they need new skills for the AI era but neither they nor their organisations can articulate what those skills are. That gap is becoming a crisis.
The Skills Gap Nobody Talks About
AI literacy training focuses on the wrong audience. Most programmes target individual contributors: how to use AI tools, how to write prompts, how to integrate AI into daily work. That is important, but it misses the bigger problem.
Leaders need different AI skills than individual contributors. They do not need to write prompts. They need to make decisions about AI strategy, manage AI-augmented teams, evaluate AI investments, and navigate the ethical and organisational implications of AI adoption.
78%
of executives say their leadership development programmes do not adequately address AI-related leadership skills
Source: Harvard Business Review Analytics, Leadership in the AI Era, 2025
Gerson's research identifies five leadership capabilities that the AI era demands:
1. Uncertainty Navigation
Traditional leadership is built on reducing uncertainty. Analyse the situation, gather data, make a decision. AI introduces a new kind of uncertainty: probabilistic tools that are usually right, sometimes wrong, and never certain. Leaders must become comfortable making decisions based on probabilistic information and managing teams that work with uncertain tools.
This is not just an intellectual shift. It is an emotional one. Leaders who built careers on being the person with the answer must learn to be the person who manages the process of finding answers in uncertain conditions.
2. Human-AI Teaming
Leading a team of humans is well-studied. Leading a team where some of the work is done by AI is not. The challenges are specific:
- How do you evaluate performance when AI handles part of the work?
- How do you develop talent when AI accelerates the learning curve for some skills and makes others obsolete?
- How do you maintain team cohesion when AI changes the nature of collaboration?
- How do you ensure accountability when decisions are partially AI-informed?
The leadership literature on team management assumes all team members are human. We need new frameworks for hybrid human-AI teams, and we need them now, not in five years when the research catches up.
Dr Gerson Tuazon
AI Strategy & Health Innovation
3. Ethical Reasoning at Speed
AI creates ethical questions faster than organisations can develop policy. A customer service AI is showing bias. A hiring AI is producing questionable recommendations. A content AI is generating outputs that might be inappropriate. Leaders need the ability to make ethical judgements about AI in real time, without waiting for a committee to convene.
This requires a different kind of ethical reasoning than traditional corporate ethics. It requires understanding AI-specific risks (bias, hallucination, data privacy, autonomy erosion) and making judgement calls that balance innovation with responsibility.
4. AI Investment Evaluation
AI investments are different from traditional technology investments. The ROI is less predictable. The timeline is longer. The compounding effects are harder to model. Leaders need frameworks for evaluating AI investments that account for these differences.
Specifically: how do you evaluate a capability that improves with use? How do you calculate ROI for an AI foundation that makes future capabilities cheaper? How do you compare the cost of AI investment against the cost of AI inaction?
Traditional business case frameworks do not handle these questions well. Leaders need updated tools.
5. Change Communication
Communicating about AI to employees, customers, and stakeholders requires a specific kind of transparency. Too much enthusiasm creates unrealistic expectations. Too much caution creates resistance. Leaders need to communicate honestly about what AI can and cannot do, what it means for jobs and workflows, and what the organisation's AI principles are.
This is harder than it sounds. Most leaders have been trained to communicate confidence and certainty. AI requires communicating uncertainty and nuance without undermining trust.
Why Current Development Fails
Gerson's assessment of current leadership development approaches:
Executive AI workshops (1-2 days) provide awareness but not capability. Leaders leave knowing that AI matters but not knowing what to do about it. Awareness without action creates anxiety, not readiness.
MBA programmes are updating curricula, but slowly. The AI modules in most business programmes focus on strategy and technology, not on the people-leadership challenges of AI adoption. And most current leaders completed their MBA before AI was a topic.
Coaching and mentoring depends on coaches who understand AI. Most executive coaches do not. They are expert in human dynamics, but the human-AI dynamics of the current moment require specific knowledge that most coaches lack.
Peer learning is useful but unstructured. Leaders learn from each other's AI experiences, but without frameworks for synthesising and applying those experiences, the learning is anecdotal and inconsistent.
What Works
From the research, three approaches show genuine results:
Immersive AI Experience
Leaders who spend a concentrated period (two to four weeks) personally using AI tools in their actual work develop materially better AI leadership capabilities than those who attend workshops. The experience of wrestling with AI's strengths and limitations, in the context of real decisions, builds intuition that no presentation can provide.
Cross-Functional AI Projects
Leaders who co-lead an AI project with a technical counterpart develop both the technical understanding and the change management skills needed for AI leadership. The project provides a concrete context for learning, and the cross-functional pairing ensures both strategic and technical perspectives are present.
Structured Reflection
Leaders who maintain a structured practice of reflecting on their AI decisions, what worked, what did not, what they would do differently, develop faster than those who do not. Gerson's research suggests a weekly fifteen-minute reflection practice produces measurable skill development within three months.
The leaders who are navigating the AI transition best are not the ones with the most AI knowledge. They are the ones with the most honest relationship with their own uncertainty.
Tim Hatherley-Greene
Chief Operating Officer
The Organisational Response
Organisations that are addressing this well share common patterns:
They invest in AI leadership development as seriously as they invest in AI technology. If your AI technology budget is ten times your AI leadership development budget, you have a problem.
They create safe spaces for leaders to experiment and fail. AI leadership skills develop through practice, and practice involves mistakes. Leaders need permission to make AI-related mistakes without career risk.
They pair AI-literate leaders with AI-developing leaders. Mentoring relationships where experienced AI leaders guide developing ones accelerate skill development across the organisation.
They update performance frameworks. If leadership performance is still measured on pre-AI criteria, leaders have no incentive to develop AI skills. The evaluation frameworks must change to reflect the capabilities that now matter.
The leadership development crisis is real, but it is solvable. Not with more AI training, but with fundamentally updated approaches to how we develop leaders for a world where AI is a constant presence in every decision, every team, and every strategy.