M.O.O.N. – My Opposite of Now

Tap the yellow “EN” or “DE” button at the top to switch to English or German. Choose one of the three yellow entry points. As soon as a message box appears, you can speak or write in any other language. M.O.O.N. will adapt to your language.

Español

Elige una de las tres opciones amarillas. En cuanto aparezca un cuadro de texto, podrás hablar o escribir también en español. M.O.O.N. se adaptará.

Français

Choisis l’une des trois entrées jaunes. Dès qu’une fenêtre de message s’affiche, tu peux parler ou écrire en français. M.O.O.N. s’adaptera automatiquement.

Italiano

Scegli uno dei tre accessi gialli. Non appena compare una casella di messaggio, puoi parlare o scrivere anche in italiano. M.O.O.N. si adatterà automaticamente.

Iwrith

בחר באחת משלוש הכניסות הצהובות. ברגע שיופיע חלון הודעה, תוכל לדבר או לכתוב בעברית.
.M.O.O.N יתאים את עצמו אוטומטית.

Nederlands

Kies een van de drie gele toegangen. Zodra er een berichtveld verschijnt, kun je ook in het Nederlands praten of schrijven. M.O.O.N. past zich aan.

Russisch

Выбери один из трёх жёлтых входов. Как только появится окно сообщения, ты сможешь говорить или писать по-русски. M.O.O.N. автоматически адаптируется.

Ukrainisch

Обери один із трьох жовтих входів. Як тільки з’явиться вікно повідомлення, спілкуйся або пиши українською мовою.

The exhibition aims to encourage people to counter an often troubling present with visions of their own. If the point is to think for ourselves, what’s the point of AI? Wouldn’t the opposite of now actually be no AI at all? 

Rather than longing for a time without AI, we set out to think about and experiment with AI differently, outside the familiar patterns: not as a means of conveying knowledge or as an information tool, but as a creative medium for the imagination. The interaction produces an unexpected kind of text.

Our AI application isn’t an answer machine that does the thinking for you. Instead, it is designed as a tool for reflection and a sparring partner that can open up new ways of thinking. The balance between serious reflection and playful absurdity is intended to create a space where you dare to think in radically different ways.

Neither the artworks in the exhibition nor this AI application provide a ready-made guide for this. Instead, they create spaces where you can discover for yourself what it means to think differently, see differently, and act differently. Will it work? Will it make a difference that matters to you – enough, for example, to justify the electricity used by this AI application? In the end, only you can decide.

Interacting with our AI is not a substitute for talking with those around you. On the contrary: why not talk to others about your experience chatting with M.O.O.N. and the result it created for you? Or take a look at the results shared by others.

At the end of the exhibition, we plan to analyze the shared texts: What are the similarities? What hopes, fears, and longings recur? 

M.O.O.N. combines a language model with selected exhibition content. This process is called retrieval-augmented generation (RAG): the European large language model (LLM) Mistral Medium 3.5 (if this model becomes unavailable, the application will use a locally hosted Mistral Small model, or Claude Haiku as a secondary fallback) was combined with a curated collection of texts on the artworks, including curatorial essays and philosophical texts related to the exhibition (see bibliography below). This keeps the AI’s responses closely focused on the exhibition. 
During development, we also extensively tested guardrails designed to prevent the generation of anti-Semitic and other hateful content.

How does the conversation work?

Visitors interact through a web-based chatbot interface, using either voice input or the keyboard. Prompts structure the conversations into five phases: opening up, focusing, disrupting, shifting perspective, finding a form. An initial system prompt also sets the tone for the AI, for example: “You are smart, concise, slightly tongue-in-cheek, surprising, never patronizing, never cynical. You tend to ask questions rather than give long explanations.” We are making the source code and the prompts used available on our GitHub account.

As it interacts with visitors, the AI identifies patterns such as recurring themes, imagery, and turns of phrase. M.O.O.N. uses these patterns to generate a personalized text. A random generator selects one of 26 predefined formats, such as a screenplay, manifesto, or dictionary entry. 

Visitors can access the tool in the exhibition at a media station or on their own smartphones. It is designed to be accessible. The language models provide multilingual support for the dialog.

Are chats stored?

The system was developed in line with the principles of privacy by design: chat data is not stored beyond your browser session. Generated texts are saved only if you explicitly consent to their publication by clicking the “Share” button. Posts are anonymous and identified only by the date and time. Chat histories are not used to train any LLM. More detailed information on data processing and privacy can be found below in our Privacy Policy.

Can AI make mistakes?

Despite its curated knowledge base and tested safeguards, M.O.O.N. may still provide inaccurate or inappropriate answers. Our system also relies on an LLM whose training data may encode certain biases. If you encounter inappropriate or offensive content, please report it so we can refine the knowledge base and/or guardrails. 

Go to the feedback form

The AI application operates as a closed system. Alongside the museum’s own exhibition materials, such as the curators’ concept, object labels, artwork descriptions and floor plan, and publicly available information, such as Wikipedia articles about the artists, it draws on the following texts:

• Alechem, Schalom, Die erste jüdische Republik, translated into German by Stefania Goldenring, Berlin 1919.
• Arns, Inke: The Big Ear; or, How to Hack the “Doomsday Radio Station”, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 226–229.
• Benjamin, Walter, Über den Begriff der Geschichte (On the Concept of History), 1940.
• Bloch, Ernst/Adorno, Theodor W.: Etwas fehlt ... Über die Widersprüche der utopischen Sehnsucht, in: Rainer Traub/Harald Wiesner (Hg.): Gespräche mit Ernst Bloch, Frankfurt 1975, p. 58–77.
• Bloch, Ernst: Excerpts from the Foreword, in Ernst Bloch: Das Prinzip Hoffnung (The Principle of Hope), Frankfurt 1959, p. 11–12 and p. 15.
• Diederichsen, Diedrich: On the Far Side of Ninety, on the Far Side of the Solar System, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 101–104.
• Edelsztein, Sergio: Circular Time, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 184–186.
• Einstein, Albert, Mein Glaubensbekenntnis (My Credo), 1932.
• Forsythe, William: Undoing, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 112.
• Friedrich, Julia: Connected Along Sharp Edges, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 264–267.
• Friedrich, Julia/Harten, Shelley: Introduction, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 28–38.
• Harten, Shelley: The Opposite of Now Is Here: Thoughts on Location, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 82–89.
• Kavelina Dana/David Riff: Undoing the Damage, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 122–127.
• Kluge, Alexander: How Walter Benjamin Explained the Weak Messianic Power to Bertolt Brecht, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 155.
• Kramer, Mario: Productive Powers, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 114–116.
• Lapidot, Elad: Sparks of Return, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 70–81.
• Menke, Christoph: Yes, but ..., in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 172–175.
• Seroussi, Benjamin: Yael Bartana’s Factory of Worlds, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 196–199.
• Sheleff, Maayan: A Polyphonic Anthem for a Fragmented Utopia, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 140–143.
• Vogl, Joseph: Unlocking the Closedness of the Past, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 156–158.
• Wajsbrot, Cécile: “In-Between”, in: Julia Friedrich/Shelley Harten (ed.), The Opposite of Now: Artistic Paths to a Different Present, Munich 2026, p. 248–250.
• Zamenhof, Ludwik Lejzer, La Espero, 1980, available online at https://de.wikipedia.org/wiki/La_Espero.