Accidental Archives

Accidental Archives examines bodies of evidence that become available for research by routes never intended to create an archive. Rescued websites, data dumps, leaks, hacks, and disclosures arising from legal proceedings provide different forms of access to an information environment whose ordinary workings are largely hidden from public view. The project reconstructs how such evidence was selected, extracted, preserved, disclosed, and transmitted before asking what it can reveal.

Case studies range from the rescued GeoCities archive to data exposed by leaks and hacks and records emerging from litigation involving major technology companies. These materials offer unusual glimpses into how the immense digital record of everyday life is produced and processed, encompassing information people deliberately share as well as photographs, interactions, behavioural traces, and other data generated or collected in the ordinary use of cloud-based services. Legal and technical records can reveal how server-side machine learning derives information from these materials, how algorithms mediate subsequent interactions, and how increasingly elaborate computational representations of users emerge from data produced at Zettabyte scale. Such disclosures make aspects of contemporary information systems available for scrutiny that their ordinary operation rarely exposes.