The Psychology of AI Adoption: Why Most Organisations Are Stuck in Stage 2
Anthropologist Kalervo Oberg's culture shock model - honeymoon, frustration, adjustment, adaptation - maps the AI adoption journey with unusual accuracy. Most organisations are stuck in Stage 2. Here is what that stage requires.
The most useful framework for understanding what organisations are going through with AI right now is not a technology framework. It is a psychology framework from 1960. Kalervo Oberg's culture shock model maps the AI transition in organisations with unusual accuracy - and explains why the path through is psychological before it is technical.
The challenge of AI adoption is most commonly framed as a technology challenge. The wrong tools. The wrong integration. The wrong data infrastructure.
But organisations that have the right tools and the right infrastructure are also struggling. The technology is largely not the problem.
The problem is the human transition - the fear, the adjustment of identity and role, and the deep uncertainty about what expertise is worth in a world where a machine can do what took years to learn.
That is a psychology problem. And a framework from classical anthropology describes it better than anything written specifically about AI.
Oberg's Culture Shock Model - Applied to AI
Kalervo Oberg was an anthropologist who studied what happens when people enter genuinely unfamiliar cultures. In 1960, he identified four stages: honeymoon, frustration, adjustment, and adaptation. The pattern, he observed, was consistent across contexts and individuals.
Applied to the AI transition in organisations, Oberg's model is more accurate than most contemporary analysis.
Stage 1 - Honeymoon
The first demonstrations of AI capability are genuinely impressive. Screening time drops. Content is generated in seconds. Tasks that took hours take minutes. The emotional response in most organisations is wonder - a sense that everything is about to change for the better.
This stage is real. The capability is not imagined. But it surfaces what AI does well and temporarily obscures where it struggles.
Stage 2 - Frustration
Integration with existing systems proves harder than the demos suggested. AI produces unexpected outputs - shortlists that favour certain profiles, decisions that cannot be easily explained, patterns that surface in the data and create concern. Job security fears spread. The ROI is harder to demonstrate than expected. Trust erodes.
Most organisations in 2026 are somewhere in Stage 2. This is not a failure - it is the necessary friction of encountering reality after the promise. But it is also where the most consequential decisions get made and where the most common mistakes happen.
The Stage 2 mistake is binary: either dismiss the concerns and double down, or pull back and conclude the technology is not ready. Both miss the actual work of Stage 2 - diagnosing exactly where AI works, where it does not, and what the human layer needs to provide.
Stage 3 - Adjustment
Organisations that navigate Stage 2 well develop explicit frameworks for the human-AI boundary. They define which decisions belong to the machine and which to the human. They invest in developing the human capabilities AI cannot replicate. They build audit processes for AI outputs. They communicate honestly with their teams about what is changing and why.
This stage requires leadership commitment that goes beyond technology adoption. It is organisational change work - psychological, cultural, and structural simultaneously.
Stage 4 - Adaptation
AI becomes operational infrastructure. The anxiety of Stage 2 fades. The focus shifts from 'are we using AI correctly?' to 'are we producing the outcomes we need?' The technology is a foundation, not a conversation.
The organisations that reach Stage 4 are not the ones that adopted AI earliest. They are the ones most honest about Stage 2 and most deliberate about Stage 3.
What Classical Psychology Reveals About the Current Moment
Oberg's insight was that the frustration stage is not a sign the transition is failing. It is a sign the transition is real. The organisations experiencing Stage 2 discomfort are the ones genuinely engaging with what AI adoption requires.
The path through the AI transition is psychological before it is technical. Understanding which stage your organisation is in - and what that stage requires - is more useful than any capability benchmark.
Where classical psychology and contemporary science converge on AI is in this observation: the capability of a system and the human readiness to work with it are separate variables. Advancing one without attending to the other produces the Stage 2 trap - powerful tools, human systems not yet designed to use them well.
The framework for getting through Stage 2 is not more technology investment. It is honest diagnosis, clear communication, and deliberate development of the human judgment that AI cannot replicate.
The Practical Implication
If most organisations are in Stages 2 and 3 right now, the investment that matters most is not in more AI capability. It is in the human infrastructure that makes AI adoption durable: change management, transparent internal communication, deliberate development of judgment and relationship skills, and explicit governance of where human accountability must be maintained.
The organisations that will lead in three years are not the ones buying the most advanced tools today. They are the ones building the most adaptive human systems around those tools.
That is a psychological and organisational challenge. Oberg understood it about culture shock. It applies, in full, to the AI transition.