We mobilise people and resources to create, curate, make sense of and use knowledge to inform policymaking across Europe.
New JRC report maps how civil servants use GenAI and how administrations can move from pilots to strategy.
Generative AI has quietly become part of daily life for many EU civil servants - drafting emails, summarising reports, even flagging compliance issues. Our latest study The adoption of Generative AI in EU public administrations: exploring individual behaviours and organisational approaches suggests that while enthusiasm is high, most public administrations are still figuring out how to harness it strategically, safely and at scale.
Drawing on 31 interviews across eight case-study administrations, the study analyses GenAI adoption from two angles: how individual public servants use the technology, and how organisations are working to assimilate it.
Civil servants often experiment with GenAI informally, using commercial tools outside official channels. This "shadow AI" use reflects genuine productivity gains, while also raising risks around data protection and oversight. Employees see GenAI as a valuable cognitive assistant, but also worry about over-reliance, skill erosion and unclear rules on appropriate use. Consequently, tensions arise in their daily working life around how civil servants experience GenAI: empowerment versus dependency, innovation versus control, and exploration versus regulation.
At the organisational level, administrations are navigating the best way to assimilate GenAI into their daily operations. The challenge is not anymore to ensure a wide adoption but rather identify the best assimilation strategy to maximize the creation of public value, given the specific needs of the organisation. That is why the report proposes a four-stage model of GenAI assimilation - exploratory, operational, transformational and reflective - showing how the value generated by GenAI can span from simple awareness to governed extension of civil servant’s cognitive capacity.
The report sets out concrete recommendations for EU public administrations, organised around three priorities: building interoperable, secure GenAI infrastructure; strengthening governance and operational readiness through training and oversight bodies; and accelerating public value through structured piloting and cross-agency collaboration.
The findings feed into the wider EU policy debate shaped by the AI Continent Action Plan and the Apply AI Strategy, supporting the shift from fragmented experimentation towards more strategic, sovereign and trustworthy GenAI adoption in government.
This report builds on the JRC's broader work of the INNPULSE team on AI and digital transformation in the public sector, and is closely linked to the Public Sector Tech Watch observatory, which tracks more than 200 GenAI-related cases across European public administrations.
23 Jun 2026
