The Experiment We are All Part Of
Right now, across thousands of organisations worldwide, a massive experiment is underway. We are integrating AI into the fabric of how humans work, decide, learn, and collaborate. We are changing the fundamental nature of human labour, cognition, and social interaction in the workplace.
And we are not measuring the human side.
We track everything about the AI: its accuracy, its speed, its efficiency gains, its ROI. We have dashboards showing how many decisions it makes, how much time it saves, how much cost it cuts. But who is tracking what it is doing to us? Who is observing how it changes the way teams communicate, reshapes what people believe their work is for, affects the skills people develop or lose, transforms the informal social structures that make organisations actually function?
The uncomfortable answer: almost no one. And that is not an accident but a mandate gap.
The Pattern We Keep Repeating
Major technological shifts often produce unintended consequences that outweigh their intended benefits.
When factories introduced assembly lines, they measured productivity gains meticulously. What they did not measure, until much later, was the psychological impact of de-skilled repetitive work, the breakdown of craft communities, the loss of worker autonomy and meaning. By the time we understood those impacts, we had built an entire industrial system around them.
When organisations introduced email, they measured communication efficiency. What they did not measure was the death of deep work, the creation of always-on culture, the subtle shift from thoughtful correspondence to reactive one-liners. We only noticed after burnout became an epidemic.
The pattern is consistent in that we optimise for what we measure, and we ignore what we do not. What we ignore tends to matter most. AI is no different, except the stakes are higher, the changes are faster, and the impacts are more fundamental.
What the Dashboards Cannot See
What is happening right now in organisations implementing AI is that it is invisible to the metrics, but visible to anyone actually watching human behaviour.
The Silent Skill Decay. A team of financial analysts used to debate complex scenarios, challenge each other’s assumptions, develop sophisticated mental models together. Now they input data, review AI recommendations, and click approve. Their analytical muscles are atrophying. In three years, when the AI hallucinates or encounters a scenario outside its training, no one will remember how to think through these problems manually. The organisation will not know it has lost this capability until it desperately needs it.
The Authority Vacuum. An AI system makes hiring recommendations. Slowly, subtly, hiring managers stop trusting their own judgement. “Well, the algorithm scored her 8.2” becomes the end of the discussion. When someone wants to override the system based on nuanced human reading of a candidate, they are asked to justify why they think they know better than the data. The power to decide, and the responsibility that comes with it, has shifted to a system no one fully understands. No one planned this transfer. It just happened.
The Meaning Collapse. A graphic designer spent years developing aesthetic judgement and creative intuition. Now AI generates options in seconds. She curates and tweaks rather than creates. Her productivity metrics are excellent. Her sense of professional identity is quietly eroding. The organisation sees improved output. It does not see the crisis of purpose forming in its creative talent.
The Learning Breakdown. Junior consultants used to learn by watching senior colleagues wrestle with ambiguous problems, seeing their thinking, their false starts, their breakthroughs. Now AI provides the analysis and seniors review and present it. The apprenticeship model that transferred tacit knowledge across generations has quietly broken. In five years, the organisation will wonder why its mid-level talent cannot handle complex ambiguity. It will not connect that to the AI implemented years earlier.
The Accountability Dissolution. When an AI-assisted decision goes wrong, who is responsible? The person who clicked approve? The data scientist who built the model? The manager who deployed it? The executive who mandated its use? Everyone has plausible deniability. Accountability, that essential ingredient of healthy organisations, has become a fog.
None of this shows up in your AI performance metrics. All of it fundamentally undermines organisational health, resilience, and sustainability.
Why Existing Roles Are Missing It
You might reasonably ask: do we not already have IO Psychologists, Organisational Development practitioners, People Analytics teams, IT governance functions, Risk officers? Why are they not catching this?
The honest answer is that they could. But they are not, for a specific reason: none of them currently hold an explicit mandate to study human-AI interaction as a cultural and organisational phenomenon.
IO Psychologists are focused on individual and team performance. OD practitioners are typically engaged around planned change initiatives. People Analytics teams are measuring engagement, attrition, and productivity, but usually through the lens of existing frameworks that were not designed for this question. IT and Risk functions are monitoring technical performance, security, and compliance. Everyone is doing their job. Nobody’s job is to watch what is actually happening between the humans and the machines.
This is not a criticism of those roles. It is a structural gap. The question of what AI is doing to human capability, meaning, judgement, accountability, and social organisation simply does not sit inside any existing mandate. It falls between them.
Enter the AI Anthropologist
The AI Anthropologist is not a new academic discipline. It is a new organisational mandate, one that can and should be held by people already working inside organisations, including IO Psychologists, OD practitioners, People Analytics specialists, IT professionals, and Risk officers, provided they are trained and explicitly tasked to look at this.
What anthropologists do is observe culture from the inside, tracking the gap between formal systems and actual practice. What the organisation chart says versus how work actually happens. What the policy requires versus what people actually do. What the technology is designed for versus how humans adapt to, resist, or quietly subvert it.
In stable times, this gap is manageable. During rapid transformation, it becomes dangerous.
An AI Anthropologist would notice the team that stopped having creative debates because AI now provides “the answer.” The manager losing confidence in her own judgement. The informal knowledge-sharing practices that quietly disappeared. The new status hierarchies forming around AI fluency. The ethical shortcuts people take when AI recommendations feel official. The skills decaying faster than anyone realised.
They would ask: how has decision-making authority actually shifted, versus how we think it has shifted? What are people learning, and what are they unlearning? Where has accountability become ambiguous? What stories do people tell about AI when leaders are not in the room?
They would document the unintended consequences we can still do something about, the cultural changes we are not planning for, the human costs we are not counting.
The Choice We Are Making Right Now
The reality is: if you are implementing AI in your organisation without someone systematically studying its human and cultural impact, you are conducting an uncontrolled experiment on your people. You are making decisions about technology deployment without data on what matters most.
You would not deploy AI without monitoring its technical performance. Why would you deploy it without monitoring its human impact?
Every day we implement AI without this function is a day we are choosing blind optimisation over informed decision-making. And we know from history that by the time the human consequences become obvious, they are also deeply embedded. Much harder to untangle. Much costlier to address.
The experiment is already running. We are all participants. The AI Anthropologist is the person whose job it is to take notes on what it is actually doing to us, while we can still shape what comes next.
That is not a luxury. It is arguably the most important role in your AI governance structure that you do not yet have.
Lisa Ashton is the Managing Director at Bioss.
