AI:s spridning i banksektorn : Konsekvenser och Möjligheter

Sammanfattning

This study examines the diffusion of artificial intelligence (AI) in the banking sector with the aim of mapping which areas and tasks are affected by AI implementation and analysing the consequences for occupational roles and work processes, as well as identifying new emerging roles and regulation. The study applies a qualitative research design combining a literature review and a document analysis of publicly available material from Swedbank, Finansinspektionen and the Swedish government's AI-strategy from 2026. The empirical material is analysed through a theoretical framework drawing on the ideas and concepts of Meyer & Rowan (1977) and DiMaggio & Powell's (1983) institutional theory, Williamson (1979) & Ouchi’s (2019) transaction-cost theory and March’s (1991) intended rationality derived from decision-making theory. The findings show how AI-adoption in the banking sector is driven by a dual logic of efficiency-seeking and institutional legitimation. Routine, or rule-base tasks are being automated which is in line with transaction-cost theory predictions, while isomorphic pressures from regulators, competitors and professional norms accelerate the broader diffusion of AI as an organisational standard. The consequences for occupational roles are asymmetric and polarizing. Occupations requiring more technical competence, complex and ethically sensitive judgements are strengthened, while low-skilled positions are at risk for substitution. Estimates suggest that up to 1,2 million banking jobs are at risk by 2030. Furthermore the study identifies a significant governance gap, wherein widespread AI-adoption co-exists with an absence of formal AI strategies and guidelines internally in the majority of firms and banks, which raises the risk of what Hamdouni (2025) terms AI Ethics Washing. The study concludes that efficiency of the operation and optimization of resources for the banking sector from AI-adoption is conditional and depends not only on heavy technological investments but also on each Nations readiness for AI implementation. The Swedish banking sector is therefore dependent on the national digital infrastructural readiness, ethical regulation and institutional oversight to see financial performance and sustainable development of AI technology.