Posted by 
Every evening, the most valuable asset in a company walks out the door. Peter Drucker built his thinking about the value of knowledge work. Knowledge workers, he argued, own the means of production, because it lives in their heads and they can take it anywhere. He drew a management lesson from it: knowledge workers must be led, not managed.
For thirty years his observation held, because thinking happened in heads, notebooks, and private files.
It no longer does.
We now think inside AI tools. The drafts, dead ends, and revisions that used to evaporate are logged, stored, and readable by a model that belongs to a vendor. For the first time, the means of production stay behind when the worker goes home. That makes data ownership the question that decides where ingenuity will come from next.
Data Ownership Is Counted in Purposes, Not Petabytes
Drucker also gave us the test for what data is worth. In 1988 he wrote in Harvard Business Review that information is data endowed with relevance and purpose. Put simply, data is worth nothing until someone puts it to use. Finding new uses used to take people and weeks. AI does it for millions of records in seconds.
So, data ownership no longer means who stores your data. It means who decides what your data is used for.
That is also what ingenuity is: finding a new use for what you already have. The engineer who notices that a maintenance log predicts cancellations creates no new data. She puts existing data to a new use, and she can only do that if her company still holds the right to it.
The Data Ownership Nobody Writes Down
Company contracts cover finished work, almost never the drafts. Yet the middle is where ingenuity lives. Every strategy deck iterated inside a copilot is a record of how a company thinks, not just what it concluded. It is the most valuable data a company produces and the least protected.
Consider what happens when someone resigns. Their laptop is wiped and their badge is deactivated. Their two years of thinking inside an AI tool stays exactly where it was. Who owns it now: the employee who did the thinking, the employer who paid for it, or the vendor who stores it? Almost no company can answer, and almost no employment contract asks.
Four Purpose Rights of Data Ownership
Data ownership now comes down to four rights that decide who can put data to use.
Decide. You choose who can use your data and can take access back anytime.
Limit. Every permission comes with a set purpose and boundaries.
Prove. You can show where your data came from and that it’s yours.
Trace. You can see who accessed your data, when, and why.
This September the mathematician Tristan Buckmaster asked publicly whether OpenAI’s model had drawn on a year of drafts he and Levent Alpöge kept in Codex, a dispute I covered in Forbes. Whatever the merits, the only party able to answer the question was the vendor.
Where Ingenuity Goes When Data Ownership Pools
A handful of AI vendors now sit inside the half-finished thinking of thousands of companies at once. Even if no one ever misuses a byte, that position is extraordinary. A vendor can see which problems an entire industry is stuck on before any company in it solves them. If ingenuity is finding new uses for what you hold, the vendors hold more of everyone’s middle than anyone in history. They become the most ingenious company nobody hired.
The wrong response is to hoard. Data locked away and never put to a new use is a storage bill. A company with perfect control and no new ideas fails Drucker’s test as badly as one that signed everything away.
What to Do on Monday
Read your AI vendors’ terms for what they say about drafts and work in progress, not just outputs. Replace lists of what a vendor may not do with lists of what it may. Set a privacy window for unfinished work, the way publishers embargo a story. Add AI work history to your off-boarding checklist and decide who keeps it. And make sure your people can move their drafts to another tool freely.
The Bottom Line on Data Ownership and Ingenuity
Drucker’s knowledge worker carried the means of production home every night. AI has given us a choice he never had to consider. We can let our thinking settle on a few servers by default, or we can decide, deliberately, who owns the middle. Ingenuity will follow whichever choice we make.
About the author:
Sandy Carter is CEO of EQUS, an agentic AI company focused on privacy-based AI, identity, and trust infrastructure. A former AWS and IBM executive who scaled multi-billion-dollar businesses, she is the bestselling author of “AI First, Human Always” and founder of Unstoppable Women of AI. She has been named an Adweek Power AI 100 and NBC Top 13 AI Business Leaders.