LIMBO

We aim to create a space where people can playfully explore their most precious memories,

questioning—individually or together—whether these memories connect us through their specific, personal and manipulable content or rather through the authenticity of the emotions they convey.

Can generative AI stimulate users’ reflection on their personal memory processes?

Will users be able to open up and share their most precious memory with a machine? Can we guide the memorizing process through the use of creative (AI) tools? Will users be able to resonate with their creations? Could they find resonance in the memories shared by others?

The internet never forgets.

Using vast internet-sourced, and often controversial, data as training material, generative AIs paradoxically depicts human experience like no other tool in history. Therefore they have become the most accessible ‘memory’ of being human. Similar to an echo in a mountain valley they generate some kind of representation of the initial input, but are altered and distorted by unseen surfaces. While AI provides us fast and widespread visualization possibilities, we encourage users to critically reflect on the hallucinations this technology creates, the biases it recreates, and the similarities to our own memorizing processes.

A multisensory installation in two compartments:

Inner Limbo: This area consists of several small chambers where users can individually enter and have a natural conversation about a memory with an AI that generates a visual representation (we call it an Echo) of their memory in an intimate setting. If they like the result, they can save it and pass it on to the Outer Limbo.

Outer Limbo: This space serves as an collective archive, where all the saved echoes are stored and brought into relation to each other in two different ways: Image Cluster: An AI model trained to compare images based on semantic similarity arranges the individual Echoes into a three-dimensional map. This map reveals clusters of related memories, projected onto the walls of the Outer Limbo. Users can navigate and zoom into specific areas of the map simply by moving through the space. Memory Pond: Using the same similarity analysis, the Echoes are merged into a continuous film, almost like an endless VHS tape, displayed within a central basin in the room.

The archive expands with each new participant in the Inner Limbo, offering a space where visitors can explore and reflect on their own memories alongside those of others, fostering both personal and collective introspection.

What the experience will be about

Participants enter the installation through the Outer Limbo, where they see an collective archive generated by the users before them. They can decide to interact and influence the projections (focusing different regions), but as they don’t have the context to them yet, their attention shifts to the chambers on the side of the room –the Inner Limbos.

After entering the Inner Limbo, they are asked by an AI to choose a memory to visualize in their mind. The AI supports this process by asking dialogical questions to aid recalling the memory. The AI reacts to the answers of the

participants by not asking predefined questions but rather adapting to the conversation. From the description of the memory, the AI generates a video representation, a so called Echo, of the memory.

Participants then have several options:
- Further adjust the generations with the AI’s help.
- Accept the Echo as it is.
- Discard the draft and end the experience.

If they choose to keep the Echo, their video representation will be added to the collective memory archive. After this, they leave the Inner Limbo and return to the Outer Limbo, where their personal memory is incorporated into the projection. As they now understand how the Echoes are created, they see them in a different context. Concluding the experience, the Outer Limbo provides an opportunity to dwell, reflect, and exchange experiences with others.

Collective memory is a function of its context.

For this reason, it is important to for us to make the installation transportable in order to bring it as close as possible to the communities that are to be connected. The Outer Limbo is built using a modular and expandable truss system. Each Inner Limbo chamber is constructed with a plug-in steel frame that can be easily dismantled, transported, and reassembled. The number of Inner Limbo chambers and the diameter of the Outer Limbo can be adapted to suit the specific needs and spatial conditions of each venue.

The Space(s) The rooms represent an otherworldly, almost divine place, devoid of concrete forms and influenced by the flow of thoughts. They feature round, organic, and edgeless shapes, with no clear beginning or end. They are only illuminated by soft and atmospheric light, which directs the focus onto the projection surfaces. The light in Inner Limbo adapts to the colors of the projection and amplifies the immersion.

Outer Limbo: This space conveys a sense of openness, offering a gathering point at the memory pond in the middle where small groups can form, as well as private retreats. The visuals on the outer wall are realized via back projection hiding all the tech.

Inner Limbo: Located at the border of the Outer Limbo, the Inner Limbo consists of capsule-like cabins whose inner walls are made of a translucent fabric on which both, the projection of the memory and a play of light, can be seen. The outer walls are made of an acoustically effective and opaque fabric to protect the privacy of the visitors.

Image of Pipeline Workflow

We aim to create a
multidisciplinary team of artists from
interactive media, animation,
music, psychology and set design,
blending artistic expression with a
strong research-driven approach

the team

headshot

Raphael Tonn
Art Director

headshot

Pascal Santaella
Technical Director

headshot

Frederik Stapf
Producer

headshot

Lena Hohm
Assistant Producer

headshot

Philomena Schmidt
Architect

headshot

Leon Brückner
Musician

headshot

Clemens Kraus
Musician