Emerging technology for economic competitiveness or societal challenges? Framing purpose in Artificial Intelligence policy

Abstract

This study examines the policy discussions surrounding the purpose of the development and use of an emerging technology. It applies the two stylized technology policy frames of economic growth and societal challenges to analyse framing of one of the key emerging technologies today—Artificial Intelligence (AI). It demonstrates that recent AI policy documents include both—economic growth as well as societal challenges—frames. While AI is a novel technology, its recent policy builds on traditional ideas about the role of technology in facilitating economic growth and competitiveness supported by well-known measures such as investment in research and highly skilled workforce. Additionally, AI policy draws on more recent frame on the contribution of technology to addressing societal challenges and the Sustainable Development Goals but presents AI as a technological solution to complex societal issues. While some interest in addressing both economic and social objectives in AI policy can be observed, the policy documents remain silent about their compatibility.

Power and politics in framing bias in Artificial Intelligence policy

“…To highlight the political relevance and implications of recent discussions about AI and bias, it is important to place AI in the broader context of debates about emerging technologies (Ulnicane, 2022) characterized by radical novelty, relatively fast growth, prominent impact, and uncertainty and ambiguity (Rotolo et al, 2015). Such emerging technologies are also characterized by hype as well as positive and negative expectations that have a performative function, meaning that irrespective of their accuracy they have important implications for shaping agendas and actions (Ulnicane et al, 2022; Van Lente et al, 2013).…”

“…If prescription for addressing bias in the technical frame is a simple technological fix, then prescriptions in the case of social frame are less straightforward. We suggest that a productive way to think about bias would be to approach it as a complex social ‘wicked problem’ (Rittel & Webber, 1973) that can be addressed by a broader strategy involving diverse stakeholders and actions (Head, 2022; Ulnicane, 2022).…”

Politics and policy of Artificial Intelligence

“…Articles in this special issue demonstrate how this co‐shaping takes place at various contexts and levels from on‐demand food platforms, their couriers, regulators and consumers (Chou & Gomes, 2023) to global rankings providing scripts for AI policy (Erkkilä, 2023) and global standard setting where states with very diverse political and cultural traditions collaborate and compete in deciding rules for the development and use of AI (von Ingersleben‐Seip, 2023). Contributions also show that this co‐shaping is particularly important, as AI with its new applications developing fast can be seen as an emerging technology (Ulnicane & Aden, 2023) characterized by high uncertainty and novelty (af Malmborg, 2023). However, not everyone has equal opportunities to participate in co‐shaping AI and it can lead to reproducing existing inequalities and exclusion of disadvantaged groups (Chou & Gomes, 2023; Giest & Samuels, 2023; Rönnblom et al, 2023; Ulnicane & Aden, 2023).…”

“…While some policy frames delimit and rule out alternatives, others can expand and offer alternative representations of problems and possible actions. Ulnicane and Aden (2023) demonstrate how social frame broadens the understanding of bias in AI beyond the traditional technical frame.…”

“…Ulnicane and Aden (2023) in their work on framing bias in AI policy indicate how an alternative framing can challenge powerful interests. Examining the framing of bias, they show how technological and social frames of intersectional bias in AI policy include power aspects and suggest alternative solutions.…”