Research

Outputs

SAIF has just begun, so what you find here is relevant research from the team — the pre-project work the project builds on. SAIF's own outputs will appear as they are published, including the SAIF Foundations Report, the Futures Methods Handbook, the Epistemic Futures Atlas, and the Governance Playbook.

Profiled

• Book chapter · 2026

  A Taxonomy of AI–Environment Relationships: AI as Savior, Destroyer, Mediator, and Infiltrator

  Henrik Skaug Sætra · AI in Society, Oxford University Press

  A conceptual map of how AI relates to the natural environment, organized through four roles: savior (AI as a solution to environmental crises), destroyer (AI’s own ecological footprint and harms), mediator (AI shaping how humans perceive and relate to nature), and infiltrator (AI entering and altering ecosystems themselves). The taxonomy underpins SAIF’s effort to think about AI and the more-than-human world beyond simple cost–benefit framings.

• Journal article · 2024

  Science fiction, sustainability, and scenario use: Comprehensive scenarios for improved strategy development and innovation

  Henrik Skaug Sætra · Technovation, 132

  The methodological foundation for SAIF’s futures work and the coming Epistemic Futures Atlas. The article argues that comprehensive science fiction scenarios — richer than conventional strategy scenarios — can improve strategy development, risk assessment, and innovation, while actively creating and shaping imaginaries of more sustainable futures. Open access.

• Journal article · 2023

  The AI ESG protocol: Evaluating and disclosing the environment, social, and governance implications of artificial intelligence capabilities, assets, and activities

  Henrik Skaug Sætra · Sustainable Development, 31(2), 1027–1037

  The foundational instrument for SAIF’s assessment work. The article proposes a flexible, high-level tool for systematically evaluating and disclosing how an organization’s AI and data capabilities, assets, and activities impact sustainability — across three scopes (direct, upstream, and up/downstream) and the environmental, social, and governance dimensions. The protocol ties into the GHG Protocol, TCFD, and the wider ESG reporting landscape, and is the basis for the interactive assessment tool on this site.

Relevant research from the team

• Book · 2022

  AI for the Sustainable Development Goals

  Henrik Skaug Sætra · CRC Press

  Analyzes how artificial intelligence enables and hinders progress on the UN Sustainable Development Goals, distinguishing impacts at the micro, meso, and macro levels. The book’s guide questions and SDG-based structure inform the impact categories used in the AI ESG Protocol and in SAIF’s broader assessment framework.

• Journal article · 2021

  AI in Context and the Sustainable Development Goals: Factoring in the Unsustainability of the Sociotechnical System

  Henrik Skaug Sætra · Sustainability, 13(4), 1738

  Argues that AI’s contributions to the Sustainable Development Goals cannot be evaluated in isolation: AI is embedded in a sociotechnical system whose broader unsustainability must be factored into any honest assessment. Introduces the micro-, meso-, and macro-level analysis of AI impacts that later instruments, including the AI ESG Protocol, build on. Open access.

• Journal article · 2021

  A Framework for Evaluating and Disclosing the ESG Related Impacts of AI with the SDGs

  Henrik Skaug Sætra · Sustainability, 13(15), 8503

  Proposes using the Sustainable Development Goals as a structured basis for evaluating and disclosing the ESG-related impacts of AI — the direct predecessor of the AI ESG Protocol, connecting AI impact assessment to the established language of sustainability reporting. Open access.