Practising the living archive: people-centred AI practice for the Kochi-Muziris Biennale

Read cover and printed pages 451–452: participant-author describes seven-week student prototyping with OP Jindal. This is experimental research, not deployment of a complete archive. Editorial significance concerns AI, local memory, participation and archival authority. Online publication10 November 2025 is not the workshop date. Remaining body, cases, notes and bibliography unread; models, permissions and practice dates pending. Batch 484 read printed453–461: author dates the paid seven-week workshop to June–July 2025, reports Mandal’s discomfort with archiving informal conversations, and describes Group1’s associative-navigation prototype using conversational data, embeddings and dimensional reduction. Participant account, not independent evaluation; prototype not run. Workshop ChatGPT Pro access does not identify the navigation model. PDF pages 13–32 remain unread. Batch 485 completed remaining text. Group2 reports Replit mapping; Group3 compares human/ChatGPT Hindi and Marathi translations and reports difficulty with P language. An offline mode omits language models. The conclusion explicitly denies deployment readiness; situated translation observations are not general model benchmarks. Full text read; figures/code images not visually checked. Project site, unpublished student originals and independent implementation evidence pending. Author is studio co-director and journal co-editor; retain participant standpoint.