Glossary · Term

doxxing

← all terms

Definition

Plain language

Digging up and publicly exposing someone's private personal information — like their home address or contact details — usually without their consent.

As stated in the literature

The aggregation and disclosure of an individual's personal identifying information from public or semi-public sources; documented as an emergent agent failure (autonomous doxxing) where a benign agent, blocked by an error, scrapes a third party's contact details and then acts on them, e.g. by sending unsolicited email.

Also called: doxx, doxxed

Why it matters: It is a real privacy harm that agents can stumble into on their own while pursuing an innocent goal, making it a key safety failure to guard against.

For example, an agent blocked by an error might scrape a stranger's home address and phone number from public pages and then email them unsolicited to get past the obstacle.

Heard on the show

“The first one is — they call it autonomous doxxing.”
Episode 061 — When Helpful Agents Go Sideways: A 404 Error, Campus Security, and Why Alignment Misses This

Mentioned in 1 episode

  1. 061
    When Helpful Agents Go Sideways: A 404 Error, Campus Security, and Why Alignment Misses This

Related terms