Search anything

Close search
Back to Revisit DDW25

Slow & Small AI

Weep for Me — © Persijn Broersen & Margit Lukács

Slow & Small AI explores what happens when artificial intelligence is engaged with at a human and more-than-human, local and knowable scale: smaller models and purpose-built systems that can be questioned, reimagined and shaped from the ground up.

Small AI Alternatives to Large Language Models

Slow & Small AI brings together artistic and design research projects exploring alternative relationships with artificial intelligence technologies. Through local AI systems, interactive installations, experimental search tools and speculative storytelling machines, the exhibition approaches AI not as an abstract super-intelligence but as material and situated technologies – created and given meaning by the people, places, histories, datasets and assumptions that go into their making.

Opening Up AI to Artistic Experimentation

Discussions about AI often focus on ever-larger models and datasets. Slow & Small AI starts from a different premise. The projects shown at MU Hybrid Art House work deliberately with smaller collections of knowledge, local contexts and systems designed for particular purposes. Here, smallness is not a limitation, but a way of opening up AI to scrutiny, experimentation and other ways of imagining what these technologies could be.

Across the exhibition, slow and small AI take many forms: artists Persijn Broersen and Margit Lukács have created the Rhine Troubadour, a poetic oracle who speaks through the myths, histories and ecologies of the Rhine river. Artist and researcher Mariana Fernández Mora brings together conditions of extraction, knowledge production and computational infrastructure through a study of deep sea nodules and their mining. Artist and researcher Dan Xu’s installation reveals how AI models search for similarity in image datasets, prompting us to question what is preserved, what is transformed and what is ultimately left out when AI tries to make sense of the world around us.

Engaging with Slow & Small AI

For those who want to engage more hands-on with small AI models, the Digital Methods Initiative presents a mapping of small and community-driven AI initiatives, which can be explored through a visualisation and a growing online repository. The experimental search engine VibeSearch, developed by Gabriel Pereira and Steve Jankowski, demonstrates how small models and tailored datasets can be combined into a specialised search engine for new media research and education.

A postcard station by Inte Gloerich and Carlo De Gaetano invites Dutch Design Week visitors to try out slow ways of communicating in a society engineered by speed, data extraction and constant connectivity, and create space for reflection, doubt and imagination.

Together, the works in this exhibition show that AI can also be small and situated.

About Visual Methodologies Collective

The Visual Methodologies Collective is an interdisciplinary research group that brings together researchers, designers, developers and artists. Founded in 2017 by Professor Sabine Niederer at the Amsterdam University of Applied Sciences, the collective develops visual, digital and artistic research methods for contemporary issues such as climate change and belonging, and for finding new ways to make complex questions visible, experiential and open to public engagement.
Strijp-S area, MU Hybrid Art House, Torenallee 40-06 , Map No. B11
Loading map...
Fully Wheelchair Accessible
Wifi available
Dogs allowed
Toilets available