Identity Mining

What is broken

Personal data is extracted, combined and used to produce value, and the person who is the subject of that data is usually outside the transaction. They agreed to a terms page. They did not agree to a price, a share or an accounting.

The extraction is quiet by design. A profile is enriched, scored, segmented and sold as a dataset, or used to train a model that will be sold many times. The value shows up somewhere: in a contract, a valuation, a better product, a higher price for an advertiser. It does not show up in a statement to the person.

Consent, in this arrangement, is a permission slip rather than a bargain. It says the company may use the data; it says nothing about what the use is worth or how any of it returns. The person cannot audit the ledger because there is no ledger. They cannot decline the extraction while still using the service, because the extraction is the business model.

This is not a claim that all uses are illegitimate. It is a claim that the exchange is unaccounted. When one side can measure the value and the other cannot, the arrangement is not a market. It is a taking with paperwork.

What the requirement would take to build

Value sharing needs mechanics before it needs sentiment. The first is metering: a record of what was extracted, for which use, by which internal team or external buyer, and when. If the company cannot produce that record, it cannot share anything honestly.

The second is attribution. An extracted profile or a trained model should carry a link back to the people whose data contributed, at the level of a cohort if individual attribution would itself be a privacy problem. The link is what makes a statement possible.

The third is a ledger the subject can read: extraction events, uses, and the value assigned to each, in plain language. Not a dashboard of vague insights — a statement with entries.

The fourth is a mechanism. Revenue share, credits, a collective licence, a dividend, a lower price: the form can vary, but it has to be a real transfer with a rule, not a promise of goodwill. Where individual payment is impractical, a collective body can hold and distribute the share.

Opt-in has to be granular enough to mean something. A person should be able to allow a use, refuse a use, and leave with what has accrued. Withdrawal should stop future extraction without confiscating the past.

None of this is free to build. Metering touches every pipeline, attribution touches every model, and a ledger is a product in its own right. That is the point: if the value is real, the accounting is part of the cost of doing business.

The test: can a person see what was taken, what it earned, and what came back to them? If the answer is only a paragraph in the terms, the profile is being mined, not shared.