Skip to main navigation Skip to search Skip to main content

Fibreglass and steel: De-imagineering AI value claims for Danish SMEs

  • Technical University of Denmark

Research output: Journal Article or Conference Article in JournalJournal articleResearchpeer-review

Abstract

‘Artificial intelligence (AI) creates value’ is a phrase that circulates through Danish policy documents and consultancy reports, across conference stages and appears in opinion pieces and political speeches. Through ethnographic fieldwork with small- and medium-sized enterprises (SMEs) and AI experts involved in AI pilot projects, as well as observations at AI promotional events, we trace how SMEs are enrolled in AI adoption as vehicles for accessing this value on a national level. Drawing on anthropological theories of value, we show that there are in fact multiple meanings of value in operation: value for society, value as revenue, value as cost-saving.We build on the concept of imagineering to describe the affective process by which promotional actors semantically transmute these forms of value, thereby enabling a neoliberal stabilisation of ‘societal
value’ as a reduction in public-sector employment. Therefore, we propose the concept of de-imagineering as an analytic that helps us unblur the semantic slippage of what AI might, can and will do, and, not least, for whom it will do those things. As an analytic device, we argue that de-imagineering enables us to challenge deterministic narratives of the value accruing from technology more broadly, and thus shift the focal plane towards those who stand to benefit from this affective promotion.
Original languageEnglish
JournalAnthropological Theory
Pages (from-to)1-21
Number of pages21
ISSN1741-2641
Publication statusPublished - 2025

Keywords

  • AI
  • value
  • hype
  • de-imagineering
  • digitalisation policy
  • ethnography
  • affect
  • imagination

Fingerprint

Dive into the research topics of 'Fibreglass and steel: De-imagineering AI value claims for Danish SMEs'. Together they form a unique fingerprint.

Cite this