Why Organizations Fear AI: The Hidden Power Struggle Behind Workplace AI Adoption
Why Organizations Fear AI: The Hidden Power Struggle Behind Workplace AI Adoption
When Dartmouth University provost Santiago Schnell resigned last week after a faculty review found he had used AI to substantially edit his published research, the story was widely covered as a tale of academic integrity. That framing is accurate as far as it goes. He used AI without disclosing it, in journals that required disclosure. The rules were clear. He did not follow them.
But zoom out one step and the story looks different. Dartmouth president Sian Leah Beilock had publicly called for "more AI" at the university. The same institution that publicly champions AI adoption just performed what looks, from the outside, like a public execution over AI use. If you want a single image that captures how organizations actually relate to AI, that contradiction is it.
The bigger story is not about one provost at one Ivy League school. It is about a dynamic playing out inside every organization that has touched AI — and most of them have.
A Tool That Cuts Both Ways
Organizations do not hate AI. They love what AI does for the work. A research paper gets tighter. A product ships faster. An analyst produces more in a week than she used to produce in a month. These are real gains, and most organizations want them badly enough to pay for enterprise licenses and run internal training sessions and put AI on the agenda at every all-hands meeting.
What organizations do not love is what AI does to the people doing the work. That is a different thing entirely.
AI compresses professional gaps. It compresses the distance between the person who spent twenty years accumulating expertise and the person who walked in the door last month. It compresses the distance between the well-connected insider who always knows who to call and the outsider who just found a better way to get the answer in an afternoon. It does this quietly, efficiently, and without asking anyone's permission.
That compression is a threat. Not to the organization as an abstraction — to the specific people inside it who built careers on being the ones who know.
The Hierarchy's Immune System
Think about what happens when a manager two levels below you starts producing at your level — or above it — because she has figured out how to use AI effectively. Nothing in the org chart has changed. The reporting structure is identical. But the practical reality has shifted in a way that makes some people deeply uncomfortable.
The reaction is not always conscious. But it shows up in policy. In the sudden interest in "responsible AI governance." In usage limits, mandatory disclosure requirements, model restrictions, and approval workflows that add friction without adding value. None of these policies are announced as protection for incumbents. They do not have to be. The result is the same.
This is the organizational immune system at work. It recognizes AI not primarily as a productivity tool but as a power redistribution mechanism. And it responds the way immune systems always respond to a perceived threat.
AI is welcome exactly as far as it serves the hierarchy. It becomes a problem the moment it disrupts it.
The Solo Sport Problem
For people working independently — consultants, freelancers, solo operators, home-based entrepreneurs — AI is pure upside. There are no politics. No optics. No one tracking how you got from A to B. You just get there faster and your clients get better work. The tool rewards ambition and competence directly, with no organizational filter in between.
Inside a company, every one of those advantages becomes a potential liability. Moving faster makes you visible. Being more productive raises questions. Getting better results puts you in quiet competition with people who have been getting acceptable results for years and would very much like to keep doing so.
That is not a cynical reading of corporate culture. It is an accurate one. Organizations run on consensus and relationships as much as they run on output. Someone who uses AI to break away from the pack disrupts the social contract of the workplace, even when the work is genuinely better. Tribes notice. Tribes respond.
The Reckoning Coming
Organizations that figure out how to live with what AI actually does — not just what it produces, but what it redistributes — will outlast the ones that do not.
That means accepting that a junior analyst who uses AI effectively should move faster than a senior analyst who does not. It means accepting that expertise accumulated over twenty years is still valuable, but it is no longer a monopoly on being right. It means building cultures where the tool is neutral and the output is what gets evaluated.
Most organizations are not there yet. Most are still trying to capture the productivity while managing the redistribution. That strategy has a shelf life. The competitor who is not managing those contradictions is getting faster every quarter.
The Dartmouth situation is a small, local version of a much larger collision. An institution that calls itself the birthplace of artificial intelligence could not figure out what to do when one of its own leaders used it. That is not a Dartmouth problem. It is the problem every organization is about to face, if it is not facing it already. The ones that solve it first will not be the ones with the best AI policy. They will be the ones that stopped being afraid of what AI does to the org chart.
Aaron Rose is the founder and publisher of Tech Reader Magazine, covering system architecture, cloud platforms, and AI policy.