AI isn’t the next internet. It’s a new layer of work. (Part 2)

In Part 1 I argued that the internet made organisations faster, while AI may make them fundamentally different because it changes who does the thinking. In this second part, I explore the implications, not just for process, but for economics, power, identity and responsibility.

Once machines can do meaningful cognitive work, the implications are not limited to productivity gains. The impact is also felt across economics, power structures, professional identity, and ethics. This is a systemic shift and treating it just like another technology rollout will often fail to meet leadership expectations.

Economics

The internet reduced transaction costs: searching, communicating, distributing. AI reduces thinking costs: analysing, drafting, synthesising, iterating. Many professional jobs are built on the assumption that cognition is scarce. Yet, if developing a solid first draft and initial analysis can be produced with minimal human intervention and consequently its cost drops drastically, the structure of work as we know it changes. Smaller teams can operate at a level that previously required layers. Some firms will become radically leaner. Others will hold on to legacy structures resulting in a cost base that becomes hard to justify.

Power

The internet didn’t threaten professional identity, it changed where and how we worked. AI redistributes the balance of where expertise resides. When a junior, who has a clear brief and good knowledge of how to work with AI and can apply sound judgement and strong prompting, can produce work at the standard of a seasoned specialist, traditional status and structures start to weaken. Authority shifts away from “who has done this for 15 years” and towards “who leverages AI and can define the problem, set direction, and evaluate outputs”. That is not a comfortable transition for leaders and organisations that rely on hierarchy for control and stability.

Psychology

Work provides identity. People adopt new tools quickly when those tools enhance their competence and support their perceived self-worth, but if those tools start raising questions about their value, resistance is very likely. AI has the potential to do just that, triggering a strong personal response such as defensiveness, minimisation, avoidance, or secret usage. That is not irrational. This is a human response to protect their identity. When you realise that what took you years to build and what can take you days to create can be replicated in seconds, you don’t just look at process, you rethink your very purpose.

Responsibility

As AI develops further, decision-making may increasingly be delegated to machines. As this happens, the ethical questions around accountability, transparency, bias and duty of care will need to be addressed. If AI is carrying out most of the cognitive heavy lifting, who is accountable? The Board and the CEO remain accountable, but as decision-making is increasingly shaped by AI outputs, the scope of that accountability will need to be redefined. What can and cannot be delegated must be clearly specified, and in all probability, leadership responsibility will become broader, sharper and more explicit.

Within this context, culture, strategy, structure and leadership all matter. It is not about hype or urgency. It is about thinking through any AI roll-out which leaders will need to approach systematically:

1. Leaders need to understand and map where in the organisation judgement happens, decisions are made, risk is held and creativity is genuinely needed. This needs to be clear because if it cannot be described, the redesign cannot happen.

2. Roles are redesigned using a hybrid approach in a planned and deliberate manner. It is not that sometimes we use AI but it is about clearly defining what AI does, what humans do, and how handoffs and accountability work in practice.

3. Governance will need to be re-engineered. The role and composition of the Board will, in many organisations, need to evolve as will the role of the CEO and the C-suite.

4. Organisations will need to manage the human impact in a responsible and humane manner. Some roles will become obsolete. Leaders will need the capability to support both those whose roles no longer exist and those whose roles change and need to fit in a redefined reality. Adoption won’t happen through policy alone, it will require honest conversations which redefine purpose, value and capabilities.

AI is not just a new channel. It changes the fundamentals. And that is why it feels like a rewrite, because, in many ways, it is.