Llum BCN is a leading lighting arts festival in Europe and a cultural event highly appreciated by the citizens of Barcelona. Organised by Barcelona City Council, Llum BCN brings together interventions by creators from all areas of the visual arts, including contemporary art, design, architecture, technology and lighting design.
During the three-day festival, the Poblenou neighborhood and the surroundings of Plaça de les Glòries became a focal point for the city’s cultural activities. Over 220,000 people enjoyed the lighting installations designed by internationally renowned artists, along with projects from the city’s top architecture and design schools.
Miralls del Demà” (Mirrors of Tomorrow) is an interactive installation for LLUM BCN 2025, developed by second-year students from the Institute for Advanced Architecture of Catalonia (IAAC), integrating expertise from the Master in Advanced Architecture (MAA), Master in Robotics and Advanced Construction (MRAC), and Master in Emergent Futures (MDEF) programs.
The installation showcased at Parc del Centre del Poblenou transforms public imagination into a collective dialogue with the future. By utilising natural language processing (NLP) and generative AI for sound-to-text and text-to-image transformation, the project invites participants to voice their thoughts, which are processed and visualised in real-time. This creates a dynamic feedback loop where individual perspectives merge into a shared, evolving narrative about what the future could be.
By combining proximity sensors, microphones, and AI-driven projections, MIRALLS DEL DEMÀ addresses critical issues such as rising CO₂ emissions, the proliferation of plastics in our oceans, and the rapid urbanisation reshaping our landscapes. The installation denounces the consequences of inaction while inspiring collective responsibility to create a sustainable future.
Making of Miralls del Demà: CONCEPT to REALITY
The installation is designed to begin with one central question about the future, answered by three different individuals. Each response, delivered through spoken word, undergoes audio filtering to detect volume and transcribe the content accurately. To ensure an inclusive and respectful public experience, content moderation automatically replaces inappropriate words with emojis. The system then uses a Large Language Model (LLM) to enrich and bring coherence to the transcribed text, imbuing the responses with vivid detail and speculative imagination. These descriptions are then transformed into AI-generated visuals, projected onto a 6×2 m curved modular LED screen composed of 48 displays (1536×512 px).
The system runs efficiently on an Intel NUC that runs the core processes, while computationally intensive tasks are outsourced to APIs, eliminating the need for a high-end GPU. An ESP32 microcontroller communicates over local WiFi to orchestrate LED animations and provide participants with intuitive guidance through color-coded signals and countdowns. Finally, all responses and generated images are archived in a Supabase database, creating a growing repository of collective visions.
IAAC x LLUM 2025
Miralls del Demà ultimately offers more than a spectacle of light and visuals;
it serves as a living mirror of societal imagination. By connecting the public domain with the future,
it fosters a new mode of participatory foresight where imagination becomes both reflection and influence.
The installation demonstrates that the future is not a distant, abstract construct but an ongoing,
dynamic process shaped by our shared narratives and interactions today.
GENERATED IMAGES – CATALOG
