Skip to main content

LUP Student Papers

LUND UNIVERSITY LIBRARIES

Emission-Related Information for Third-Party GenAI Services

Henriksson, Edvin LU and Nilsson, Joel (2026) INFM12 20261
Department of Informatics
Abstract (Swedish)
Third-party generative AI services are increasingly embedded in organisational work, yet
their emissions-related information is difficult for user organisations to access, interpret and
govern. Scope 3-related reporting therefore raises an emerging information systems problem:
how organisations can make supplier-dependent, uncertain and technically opaque emissions
information traceable and usable. The study is based on a qualitative, exploratory research
design. Empirical material was collected through semi-structured interviews with seven
respondents from Sweden-based organisations and analysed through thematic analysis,
informed by information management, information governance and Organisational
Information Processing Theory.... (More)
Third-party generative AI services are increasingly embedded in organisational work, yet
their emissions-related information is difficult for user organisations to access, interpret and
govern. Scope 3-related reporting therefore raises an emerging information systems problem:
how organisations can make supplier-dependent, uncertain and technically opaque emissions
information traceable and usable. The study is based on a qualitative, exploratory research
design. Empirical material was collected through semi-structured interviews with seven
respondents from Sweden-based organisations and analysed through thematic analysis,
informed by information management, information governance and Organisational
Information Processing Theory. Findings show that emissions from third-party GenAI
services are generally recognised as potentially relevant, but rarely formalised as a distinct
reporting or governance object. Handling the issue depends on cross-functional coordination
between sustainability, IT, procurement, reporting and management, while supplier
dependence creates a central barrier because key data, methods and infrastructure remain
externally controlled. Existing reporting frameworks provide a broad starting point, but are
often too coarse to make AI-related emissions visible, separable or decision-useful. Overall,
the thesis argues that the issue is not primarily a carbon calculation problem, but an
organisational information governance problem. Feasible handling depends on pragmatic
traceability, transparent assumptions, supplier dialogue and responsible GenAI use. (Less)
Please use this url to cite or link to this publication:
author
Henriksson, Edvin LU and Nilsson, Joel
supervisor
organization
alternative title
A Swedish Qualitative Interview Study of an Emerging Scope 3 Issue
course
INFM12 20261
year
type
H1 - Master's Degree (One Year)
subject
keywords
Scope 3 emissions, Generative AI, Sustainability reporting, Information management, Information governance
language
English
id
9237175
date added to LUP
2026-06-16 10:24:10
date last changed
2026-06-16 10:24:10
@misc{9237175,
  abstract     = {{Third-party generative AI services are increasingly embedded in organisational work, yet
their emissions-related information is difficult for user organisations to access, interpret and
govern. Scope 3-related reporting therefore raises an emerging information systems problem:
how organisations can make supplier-dependent, uncertain and technically opaque emissions
information traceable and usable. The study is based on a qualitative, exploratory research
design. Empirical material was collected through semi-structured interviews with seven
respondents from Sweden-based organisations and analysed through thematic analysis,
informed by information management, information governance and Organisational
Information Processing Theory. Findings show that emissions from third-party GenAI
services are generally recognised as potentially relevant, but rarely formalised as a distinct
reporting or governance object. Handling the issue depends on cross-functional coordination
between sustainability, IT, procurement, reporting and management, while supplier
dependence creates a central barrier because key data, methods and infrastructure remain
externally controlled. Existing reporting frameworks provide a broad starting point, but are
often too coarse to make AI-related emissions visible, separable or decision-useful. Overall,
the thesis argues that the issue is not primarily a carbon calculation problem, but an
organisational information governance problem. Feasible handling depends on pragmatic
traceability, transparent assumptions, supplier dialogue and responsible GenAI use.}},
  author       = {{Henriksson, Edvin and Nilsson, Joel}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Emission-Related Information for Third-Party GenAI Services}},
  year         = {{2026}},
}