What is your personal data worth?

For most of the last decade, the honest answer to what is your data worth was: disappointingly little. Divide a large platform's advertising revenue by its user count and you land somewhere between a few dollars and a few tens of dollars per person per year. That number was used to end the argument. If your data is worth twelve dollars, why are you making a fuss?

The number was always the wrong measure, and AI has made it obsolete.

Why the advertising number was misleading

It measured what your data was worth to one buyer, for one purpose, at one moment. It was a price set in a market where you were not a participant. It said nothing about the value the same information could create for you if you could actually use it.

Health data is the clearest case. To an advertiser, a decade of wearable readings is worth a fraction of a cent per record. To you, in combination with the rest of your history, it might be the thing that catches something early. Same data, two valuations, orders of magnitude apart. The market price was never a measure of worth. It was a measure of who was allowed to bid.

What AI changed

The valuable asset moved. Behavioural data, the clicks and locations and purchases, was valuable because it predicted what you might buy next. It was a proxy, and a crude one.

What AI makes valuable is knowledge: what you actually know, how you explain it, how you reason through a problem in your field. That is not a proxy for anything. It is the thing itself, and it is far more concentrated. A consultant's twenty years of frameworks are not worth twelve dollars a year to anybody. They are the entire basis of a business.

This is why the ownership question got sharper rather than softer. When the asset was your clicks, losing control of it was an annoyance and a slow erosion of privacy. When the asset is your expertise, losing control of it is competitive.

Three ways knowledge produces value

• Leverage. The same explanation delivered to a hundred people at once instead of one at a time.
• Reach. Questions answered in time zones you are asleep in, by people who would otherwise have gone elsewhere.
• Continuity. Knowledge that stays accessible when the person who holds it is unavailable, on leave, or has moved on.

None of these require selling anything. That was the flaw in the data-as-a-commodity thinking of the last decade: it assumed the only way to realise value was to find a buyer. The better model is that you keep the asset and it works for you.

The condition attached

All of it depends on one thing. The value stays with you only if the knowledge stays yours.

Upload your material to a system that folds it into a shared model and you have converted a durable asset into a one-off convenience. The system gets more capable. You get a chat window, and so does everyone else, including people who would otherwise have hired you.

So the practical answer to what is your data worth is: it depends entirely on who holds it. Held by a platform, priced as advertising inventory, it is worth a few dollars a year. Held by you, and put to work representing what you know, it is worth whatever your expertise is worth.