Grief moves online through tools built for archives, likes and AI simulation.
A participatory memorial where absence is kept as structure: a growing sound spiral shaped by contributions, not algorithms.
A survey of 82 people found 57% uncomfortable with AI simulating someone they lost. The prototype was tested with 6 participants.
A participatory digital memorial built on one premise: absence should be preserved, not filled.
Where other platforms simulate the deceased through avatars and chatbots, Chorale chooses restraint: memory builds up through contribution, witnessing and silence, as a sound spiral shaped by its community.
Mourning has moved online, onto tools never built for grief: archives that freeze a profile, feeds that give loss a like count, and AI that offers to simulate the person you lost. That last idea was the one people rejected hardest, and it became the starting point.
Research question How can interaction design support collective digital remembrance without simulating the deceased, while preserving absence as a structural condition of the system?
Continuing Bonds theory framed the answer: a bond can continue as connection rather than closure. Working as research through design, the interface became the way the question was explored.
A literature review, then an exploratory survey of 82 people in Turkish and English. Each finding shaped a decision:
Exploratory survey, 82 participants, both language versions combined. Source: thesis, chapter 6.3.2.
Sound is temporal.
Sound is relational.
Sound persists without fixing identity.
Sound builds a whole out of fragments and never fully disappears; it disperses. A person's memory behaves the same way, which is why sound became the medium. A chorale, a piece sung by many voices without a lead, gave the project its name.
The spiral emerged through sketching, not as a starting point. Alongside it, names like Thrum, Choir and Echoir were tested until Chorale fit.
Wireframes worked out the logic: a code-gated entry, and a memorial page that holds both the shared spiral and one person's story, tested with real names and dates instead of placeholders.
Entry: code-gated access
Generic template
With a real memorial's details
Refined header, age from dates
No likes, no comments, no AI replica: a co-created soundscape instead of a timeline. Three roles came out of the research, and people can move between them as their grief changes.
Opens the chorale and holds governance, as a relationship rather than an algorithm.
Existing platforms: governance is reactive or opaque
Adds memories and chooses who sees each one.
50% prefer co-creating; 48% want private and shared
Takes part by listening and witnessing; nothing is required.
57% prefer passive viewing
No likes, comments or rankings. Less is protection.
EvidenceSurvey: discomfort with performative visibility. Testing: removing metrics lowered the pressure people expected.
The deceased is never represented or simulated.
Evidence57% uncomfortable with a simulation; only 5% wanted to talk to the deceased.
Memory as accumulation and echo, not a fixed archive.
Evidence61% keep a memory alive by talking with others. Continuing Bonds: connection, not closure.
Users choose how close they get.
EvidenceNo consensus on privacy; 48% wanted a mix of private and shared.
Silence counts. Nothing is ranked or required.
EvidencePassive viewing (57%) was preferred over co-creating (50%).
Moderation is care held by caretakers, not enforcement.
EvidenceExisting platforms: reactive or opaque governance. Testing: caretakers read as care, not control.
participants
age range
task flows
Face-to-face, 20-minute task sessions (entry, creating a chorale, exploring and contributing), judged on usability, metaphor and ethical legibility, then SUS and UEQ. The spiral read as accumulation and removing metrics eased performative pressure. SUS was high; UEQ was positive on all six dimensions, with Novelty more varied, as a restrained interface would expect.
Friction did not emerge from complexity, but from the intersection of participation and governance.
Symbols now come from contributions, not templates.
EvidenceParticipants hesitated on these options; 51% of the survey was unsure about traditions.
Private remembrance and governance shouldn't share a navigation level.
EvidenceSeveral participants questioned why they sat side by side.
Stated where they apply, not explained afterwards.
EvidenceTwo of six participants asked what the role could do.
Not only on the final settings page.
EvidenceDirect participant suggestion.
Six participants aged 24–30 make the findings directional, and long-term governance across changing caretakers remains open. Next: a coded version with dynamic zoom and real-time sound, longitudinal studies, and cross-cultural testing.
Absence is not corrected. It is preserved.
Continuity emerges through contribution, not simulation.
Memory continues.