Work support tools built on generative artificial intelligence (AI), like Microsoft 365 Copilot and other AI assistants, are increasingly the object of experimentation in public sector. Yet, empirical research on workers' experiences with these tools remains limited. This article draws on 2.5 years of ethnographic fieldwork on AI innovation in two Finnish organizations and three cases of such tools to critically explore the positioning of users as ‘moral crumple zones’ who bear the burden of ensuring the effective and ethical use of these emerging tools. The findings capture five mechanisms that produce such zones in the use of generative AI tools for administrative and frontline work: (1) the assumption of human oversight and liability, (2) the immaturity of tested tools marketed as ‘sparring buddies’, (3) trial-and-error learning offloaded to users, (4) limited transparency of AI tools' development and (5) persuasive fallibility and the burden of vigilance placed on users. These challenges place users in roles that tend to shield the tools from broader scrutiny, meaning that users effectively absorb the burden of ethical and operational oversight and liability. This article contributes to understanding the often-thorny challenges around the uptake of generative AI tools in public sector and emphasizes the need for stronger accountability pathways—not only for users but also for other actors shaping these tools and their deployment. • AI-based tools are tested with promises to enhance work efficiency in public sector. • The proposed framework helps capture five mechanisms causing ‘moral crumple zones’. • AI-based tools' testers face epistemic, interpretative and operational opacity. • The findings reveal a misalignment between testers' control and accountability. • The study calls for real accountability pathways that go beyond individual users.
Marta Choroszewicz (Thu,) studied this question.