The leadership reset: Thriving in the age of AI
Dr. Narendra Laljani, Executive Fellow and Programme Director, explores the merits of a mindset shift from authority to ambidexterity in the age of AI.
In 1970, futurist Alvin Toffler warned in Future Shock that accelerating rates of change would leave us overwhelmed and disoriented, and unsure how to respond. That future has arrived.
What we are living through now is not simply another wave of innovation. AI is reshaping how decisions are made, how work is organised, and what it means to lead. The pace is relentless, the implications are uneven, and, crucially, there is no definitive playbook.
This creates a tension at the heart of modern leadership, where many of the instincts that built successful careers are now the things that can hold leaders back.
The leadership hangover
We like to think leadership practices have moved on. But in reality, many organisations are still anchored in a set of outdated assumptions which may be described as a ‘leadership hangover’. That leaders should have the answers. That leaders focus on the ‘big picture’ while others on the operational detail. That authority comes from position. That leadership is an individual act. That strength means certainty.
These ideas were always questionable, but in an AI-driven world, they are actively unhelpful. Andy Grove, the co-founder of Intel, once observed that “only the paranoid survive”. Today, that paranoia is less about competitors and more about the risk of becoming obsolete, by relying on models of leadership that no longer fit the environment.
There is no steady state
A common question asked by organisations today is “what does the end state of AI look like?”. And whilst an obvious question to pose, it’s the wrong one. AI is not a transformation with a clear destination. It is an ongoing condition which means that leadership cannot be episodic. It must be continuous, adaptive, and at times uncomfortable.
There will never be a point at which an organisation can say “we have arrived”. The more useful question is “how well can we operate without a fixed endpoint?”.