Polylocality....at Pontevedra

At the end of February 2026, artists, researchers and technologists, including our lead researcher - Ayodele Arigbabu, met at Universidade de Vigo’s Campus CREA in Pontevedra for the 2nd General Meeting of the COST Action “Artistic Intelligence — Responsiveness, Accessibility, Responsibility, Equity” (ARTinRARE). 

___________________________________________

The digest below draws on and adapts material from the “Report to the Action” prepared by Jacqueline Rowley (Vice Chair, Collective Intelligences working group) for the 2nd General Meeting of COST Action CA23158 (ARTinRARE).

___________________________________________

Rethinking AI Through Artistic Practice

At the second general meeting of the ARTinRARE network - the European initiative that explores how artistic intelligence can shape responses to digital technology - the Collective Intelligences working group hosted a session to better understand how this kind of knowledge actually operates. Rather than drafting blueprints for AI governance, the group focused on finding a shared vocabulary to describe how artists build collective intelligence across multiple, highly specific local situations.

The starting point was a keynote address titled “Local, Polylocal”, by Michael Schwab, editor‑in‑chief of the Journal for Artistic Research, who introduced the concept of polylocality. He argued that knowledge does not exist in a single, universal space. It forms in dense, specific situations called "locals." These locals interact and negotiate with one another rather than merging into a single system. The group adopted this idea to see if it could help them understand how artistic research operates differently from standard AI models which often flatten context to extract data. Artistic practice on the other hand relies on keeping that context intact.

Schwab suggested that instead of starting with AI as a technology, it might be more useful to start with “locals”: the concrete situations where knowledge is dense and context matters, such as a particular rehearsal room, a research site in a forest, or an online community with its own codes. He argued that many such locals exist side by side, and that their relationships are negotiated rather than fixed. He called this polylocality.

For the Collective Intelligences group, this was more than an abstract idea. It offered a way to describe what many artistic projects already do: hold together different perspectives, bodies, places and tools without forcing them into a single story. The workshop by the working group therefore set out to see whether polylocality could become a shared lens for future work.

What participants brought into the room

The session began with a short reminder of themes from an earlier meeting in Porto about audiences and how people are affected by artistic work. Each participant then introduced their own practice in relation to polylocality. This produced a series of snapshots rather than a formal round of minutes.

Some spoke about the material design of data structures and how those choices shape what counts as knowledge. Others described multispecies performances where human and non‑human bodies share space, or research in fire‑prone landscapes that connects ecology with questions of justice and land use. There were feminist frameworks that talked about “hyper‑polylocality” and “hyper‑polyphony”, studies of swarm behaviour and social networks, experiments with embodied AI, and the suggestion that AI might be approached as a form of literature, something that needs careful reading from different positions.

Across these accounts, a common concern emerged: how to keep context intact when technical systems tend to strip it away in the name of scale and speed.

Two paths of discussion

The workshop then split into two groups. One stayed close to Schwab’s keynote and asked why polylocality should be a starting point at all. They tested its strengths and limits, asking whether it clarifies ongoing work or risks becoming another fashionable term.

The second group took a more practical route. Rather than argue about definitions, they began collecting verbs that describe what has to happen when a local situation opens into a wider field: listening, attuning, exchanging, crossing, improvising, repeating, experimenting, interpreting, commoning, resisting, holding space, inhabiting, opening, mythmaking, coining new terms, tracing, decontextualising and recontextualising. They treated these verbs as tools that can prompt actual actions in research and cultural projects.

When everyone came back together, Schwab highlighted “connectors” as a way of thinking about the links between locals that do not erase their differences. He asked whether the group was uncovering something new or finally naming what artistic practices have long done. The group agreed that the question should stay open, as a way to keep their work alert.

Verbs instead of static definitions

The verb‑based approach became one of the most concrete outputs of the meeting. Instead of producing a fixed list of concepts, the group organised the verbs into clusters that correspond to different aspects of artistic intelligence.

One cluster centres on reception and attunement, making listening, curiosity and sensory attention basic conditions for any meaningful exchange. Another cluster focuses on movement and exchange, describing how ideas and practices travel between contexts while trying to keep their specificities intact. A third cluster captures the iterative character of much artistic work, where improvisation and repetition build understanding over time. A fourth speaks directly to the politics of knowledge, with commoning, resisting and holding space pointing to issues such as ownership, access and extraction. The final cluster deals with language and world‑making, acknowledging that inventing or adjusting terms is itself a way of shaping what can be imagined.

In his closing remarks, Florian Schneider described this way of working as a continual practice of taking things out of context and placing them back differently. In that light, polylocality appears as an ongoing process rather than a finished framework.

Links to broader aims

This verb‑based vocabulary is not just a local experiment from one workshop. It feeds into a shared “Artistic Intelligence Reference Framework” that the Action is developing to help connect very different artistic research projects without flattening their distinctions. It also touches the concerns of other groups within ARTinRARE, such as adaptation strategies for cultural sectors facing digital change, and policy discussions around copyright, open‑source sustainability and the ethics of large‑scale data use.

By grounding general questions about AI and collective intelligence in specific practices and verbs, the Pontevedra meeting offered a way for an international network to think together without erasing the local conditions where their work actually takes place.