Skip to content

Data-subject request ​

A person's request to an organization about their own personal data, such as a copy of it or its deletion, under privacy laws like the GDPR.

The "data subject" is the person the data is about. In the GDPR, the best-known requests are access (Article 15: confirmation of whether their data is processed, and "a copy of the personal data undergoing processing") and erasure (Article 17, the right to erasureRight to erasure: The GDPR right (Article 17) to have one's personal data deleted on certain grounds, with exceptions, for example data needed for legal claims.). Other privacy laws have similar rights under other names. DSR and DSAR (data-subject access request) are common short forms.

Article 12 sets the ground rules for requests under Articles 15 to 22: answer "without undue delay and in any event within one month of receipt", extendable by two further months for complex or numerous requests; free of charge as a rule; and, for Articles 15 to 21, if there is reasonable doubt about who is asking, the organization may ask for the information needed to confirm their identity.

For an AI application, the hard part is finding the data. A person's details end up in run inputs, model answers, tool results, agent memory, approval requests, test fixtures and logs. Searching every row for an email is slow and misses runs that held only a customer id. Recording at write time which person a run was about (a subject referenceSubject reference: A value saved with a run that identifies the person the run was about, so that person's runs can be found later without searching every record.) makes the later search cheap.

Example ​

A customer emails: "Send me everything you hold about me, then delete it." The privacy team confirms it is really them, then searches by their email and customer id. The export finds 14 agent runs and one memory session. Before sending it, they check it: one support run also names a second customer, whose details must not go out with it (Article 15(4) says the copy must not adversely affect others' rights). Then they run the deletion. Rows under a legal holdLegal hold: An instruction to keep certain data from being deleted, whatever its age, because it may be needed for a legal matter. stay, and the reply says so.

Common confusion ​

  • Access and erasure are different requests with different exceptions. An access request never deletes anything.
  • Deleting the account is not erasure. The person's data also sits in runs, memory, caches and backups.
  • A request is not proof of identity. Handing one person's data to someone else is a breach in its own right.

In L2M ​

Data-subject requests are optional and off until used. An admin, or someone with a deployment-wide role granting data_subject:manage, enters the person's value; L2M follows optional per-workflow subject references (stored as keyed hashes) and can also search stored data. It builds a zip export that downloads once, or, after a preview is confirmed, deletes what it found, skipping rows a legal hold keeps. Each request has a due date (SUBJECT_REQUEST_DUE_DAYS, 30 by default) and an overdue notice. The Settings table lists the limits.

Further reading ​

This website (not the L2M product) uses Google Analytics, only after you accept cookies. Cookie settings