Finding connections between video-art archives

Two video-art collections are becoming a test of how AI connects archival evidence. LI-MA describes ArcaiArt workshops comparing its Amsterdam collection with REWIND Artists’ Video at the University of Dundee. The questions concern diverse scans, recurring information and the narratives AI might construct from them.[1]

The materials extend well beyond moving images. REWIND’s database includes clips, interviews with artists and critics, scanned articles, essays and exhibition ephemera. Its research began with British video art of the 1970s and 1980s. These records offer different ways into a work: an interview records someone’s account, while an exhibition document places it within an event.[3]

LI-MA’s collection draws on organisations including MonteVideo, Time Based Arts and de Appel. Its distribution work prepares archival and presentation files, stills, excerpts and promotional material. That adds a practical dimension to interpretation: a record may have been made to circulate a work, document a performance or prepare a screening. Its purpose matters when reading what it says.[4]

Read together, these collection descriptions suggest a practical question: can an AI-generated connection be traced back to the interview or exhibition document that supports it?[1][3][4]

The National Archives dates ArcaiArt to 2025–2028. The partner account describes research into how AI interprets and personalises responses; it supplies no comparative performance results. The promise being investigated is a more useful route through archival material.[1][2]

Report an error or suggest a correction