A tool that reads Environmental Impact Assessments through the eyes of the living world.

Developed as part of the Sentient Futures Project Incubator, March–May 2026, under the mentorship of Elahe Rajabiani.

πŸŽ₯ Video: Here πŸ”— Website: Coming Soon

πŸ“¬ Get in touch: We'd love to hear from you β€” whether you work with EIAs, think about more-than-human futures, or want to collaborate on what comes next. If you want to access the demo once we publish it, drop your contact info: Leave your details here


About the project

The problem

Environmental Impact Assessments (EIAs) β€” the EU's primary instrument for evaluating a proposed project's effect on nature β€” are written by humans, for humans. The ecosystems, species, and communities of life that will live with the consequences of the project have no voice in the process. The frameworks used by EIAs are skewed toward instrumental valuations of more-than-human life, they rely almost exclusively on natural scientific literature, and operate on timescales far shorter than the ecological processes they affect. Most of the life implicated in these decisions is simply not in the room.

Our solution

The More-than-Human Policy Reader analyses EIAs through six structured lenses β€” values of nature, bodies of knowledge, temporal and spatial scale, webs of interdependency, cultural memory, and relational reframing β€” cross-referenced against live biodiversity and citizen science records, species interaction data, and cultural heritage archives. The result is a fuller, multispecies picture made visible to decision-makers: it is not a replacement for expert assessment, but a different lens with which to assess the consequences of a project β€” one that listens for what is often left out of the frame.


How it works


The Reader is built as an LLM pipeline connected to four external databases: iNaturalist and GBIF for species occurrence data, GloBI for ecological relationship networks (who pollinates what, who depends on whom), and Europeana for cultural memories of the landscapes in question. A user uploads a PDF EIA which triggers the analysis across the six lenses, each producing confidence-rated outputs that reflect how densely a claim is supported within the source document. The interface surfaces an ecological map of the project site, knowledge system gaps, a timescape visualisation placing the project alongside ecological and geological time, a web of interdependencies, cultural artefacts, and a rewritten problem statement through a more-than-human relational frame.

Lens Framework/Approach What it reveals Why it matters
Values of Nature IPBES framework Whether nature appears only as a resource, or whether relational and intrinsic values are also present Most EIAs are dominated by instrumental value β€” making this bias visible is the first step to challenging it
Bodies of Knowledge Epistemological diversity Which knowledge traditions are cited β€” and which are absent Scientific literature alone misses place-based and traditional knowledge that often captures what formal data cannot
Time & Space Temporal and spatial scales Whether the project's timescales match the rhythms of the living systems it affects Policy cycles run in years; ecological processes run in centuries β€” decisions made at the wrong timescale cause harm that outlasts the project itself
Web of Interdependencies Ecological network analysis Which species are named, which are missing, and how they depend on one another A decision that disrupts one relationship can cascade through an entire ecosystem in ways no single-species assessment would catch
Cultural Memory Europeana cultural archive How this landscape has been seen, documented, and valued across centuries Economic valuation cannot capture what a place means β€” cultural memory makes legible a longer history of human relationship with living systems
Relational Reframing Indigenous ethics and community practices A rewritten problem statement centred on kinship, reciprocity, and shared responsibility The way a problem is framed determines what solutions are possible β€” a relational frame opens options that compliance language forecloses

Conceptual underpinning

Data sources