Advanced

Artificial Intelligence for Sustainability. On theoretical limitations, practical potentials and political discourses

Kettenburg, Annika LU (2019) In Master Thesis Series in Environmental Studies and Sustainability Science MESM02 20191
LUCSUS (Lund University Centre for Sustainability Studies)
Abstract
Several governments recently adopted strategies to facilitate the development of artificial intelligence (AI), one of the reasons being the pursuit of sustainability – a claim with little scientific grounding so far. In this thesis, I strived to outline the scope of AI’s potentials by showing i) the limits of its theoretical and practical usefulness and ii) the political barriers to using it for sustainability.
Through the method of immanent critique, I demonstrated that AI shares most of its theoretical limitations with statistical methods in general and concluded that it is best suited for the narrowly defined area of closed, formal systems with low decision stakes. To assess AI’s practical potentials, I conducted a systematic... (More)
Several governments recently adopted strategies to facilitate the development of artificial intelligence (AI), one of the reasons being the pursuit of sustainability – a claim with little scientific grounding so far. In this thesis, I strived to outline the scope of AI’s potentials by showing i) the limits of its theoretical and practical usefulness and ii) the political barriers to using it for sustainability.
Through the method of immanent critique, I demonstrated that AI shares most of its theoretical limitations with statistical methods in general and concluded that it is best suited for the narrowly defined area of closed, formal systems with low decision stakes. To assess AI’s practical potentials, I conducted a systematic literature review searching for AI use cases that contribute to the indicators of the German sustainability strategy. The resulting use cases mostly described how AI increased efficiencies or generated empirical knowledge, yet they only marginally addressed the major obstacles to sustainability identified by the government.
Since technological possibilities do not linearly translate to real-world adoption, I turned to the realm of politics to unveil dominant discourses that may shape the goals and areas of AI’s use in society. As an exemplary space of AI’s discursive mediation, I examined the German government’s national AI strategy using critical discourse analysis. It resulted that the government prioritized discourses of competitiveness and technological determinism over sustainability considerations. Tensions between different interest groups were linguistically concealed, while competition and technological progress were invoked as external threats to be faced in consensus. This strategy of depoliticization legitimizes and naturalizes current power relations to the advantage of certain actors. The revealed primacy of market logics further reduced the likeliness of realizing AI’s limited sustainability potentials. Yet, alternative ways to govern AI exist and allow overcoming the constructed necessity of neoliberal politics. (Less)
Please use this url to cite or link to this publication:
author
Kettenburg, Annika LU
supervisor
organization
course
MESM02 20191
year
type
H2 - Master's Degree (Two Years)
subject
keywords
artificial intelligence, immanent critique, critical discourse analysis, sustainability science, critical theory of technology
publication/series
Master Thesis Series in Environmental Studies and Sustainability Science
report number
2019:027
language
English
id
8980218
date added to LUP
2019-06-10 10:45:20
date last changed
2019-06-10 10:45:20
@misc{8980218,
  abstract     = {Several governments recently adopted strategies to facilitate the development of artificial intelligence (AI), one of the reasons being the pursuit of sustainability – a claim with little scientific grounding so far. In this thesis, I strived to outline the scope of AI’s potentials by showing i) the limits of its theoretical and practical usefulness and ii) the political barriers to using it for sustainability.
Through the method of immanent critique, I demonstrated that AI shares most of its theoretical limitations with statistical methods in general and concluded that it is best suited for the narrowly defined area of closed, formal systems with low decision stakes. To assess AI’s practical potentials, I conducted a systematic literature review searching for AI use cases that contribute to the indicators of the German sustainability strategy. The resulting use cases mostly described how AI increased efficiencies or generated empirical knowledge, yet they only marginally addressed the major obstacles to sustainability identified by the government.
Since technological possibilities do not linearly translate to real-world adoption, I turned to the realm of politics to unveil dominant discourses that may shape the goals and areas of AI’s use in society. As an exemplary space of AI’s discursive mediation, I examined the German government’s national AI strategy using critical discourse analysis. It resulted that the government prioritized discourses of competitiveness and technological determinism over sustainability considerations. Tensions between different interest groups were linguistically concealed, while competition and technological progress were invoked as external threats to be faced in consensus. This strategy of depoliticization legitimizes and naturalizes current power relations to the advantage of certain actors. The revealed primacy of market logics further reduced the likeliness of realizing AI’s limited sustainability potentials. Yet, alternative ways to govern AI exist and allow overcoming the constructed necessity of neoliberal politics.},
  author       = {Kettenburg, Annika},
  keyword      = {artificial intelligence,immanent critique,critical discourse analysis,sustainability science,critical theory of technology},
  language     = {eng},
  note         = {Student Paper},
  series       = {Master Thesis Series in Environmental Studies and Sustainability Science},
  title        = {Artificial Intelligence for Sustainability. On theoretical limitations, practical potentials and political discourses},
  year         = {2019},
}