This study explores how white-collar workers use generative artificial intelligence (GenAI) in their everyday work practices and examines the implications for learning and sustainability. While prior research has largely focused on adoption and productivity outcomes, less is known about how learning develops through continuous interaction with GenAI in real work contexts. Addressing this gap, the study adopts an interpretivist and qualitative approach, drawing on in-depth interviews, shadowing sessions, and logbooks from six employees in Swedish industrial organisations. 

The findings show that GenAI is primarily integrated as an efficiency-oriented, augmentative tool that supports existing work practices rather than transforming them. Employees use GenAI to streamline routine tasks, such as information retrieval, text generation, and coding, while maintaining active control over outputs. At the same time, learning emerges as an iterative and practice-based process embedded in daily work. Employees develop knowledge through continuous interaction with GenAI, including prompting, evaluating, and refining outputs, rather than through formal training.

Furthermore, the results reveal a tension between short-term efficiency gains and longer-term implications. While employees recognize both the benefits and risks of GenAI, including cognitive dependency and sustainability concerns, these reflections remain weakly integrated into everyday work practices. Sustainability is therefore acknowledged at a conceptual level but is not actively guiding how GenAI is used in practice.

The study contributes to the understanding of GenAI as a technology that shapes learning through everyday interaction rather than structured implementation. It highlights the importance of supporting reflective and critical use of GenAI to ensure that efficiency gains do not come at the expense of long-term learning and sustainable work practices.