Long-term real-world usage and outcomes of GhostWriter

Investigate how participants use GhostWriter or similar large language model-powered writing systems over extended periods in real-world tasks, and characterize long-term interactions and outcomes that are not observable in short, lab-based sessions.

Background

The evaluation consisted of two one-hour sessions focusing on editing and generative writing tasks. Due to this short-term setting, the study design does not capture longitudinal behaviors or real-world adoption patterns.

The authors explicitly note the inability to observe longer-term usage and provide the example of understanding how participants would use such systems over extended periods for real-world tasks, identifying this as an unresolved question.

References

We could not capture long-term interactions and outcomes with GhostWriter in just two study sessions, e.g., how participants would use such a system over a longer period to perform real-world tasks.

— GhostWriter: Augmenting Collaborative Human-AI Writing Experiences Through Personalization and Agency  (2402.08855 - Yeh et al., 2024) in Section 7: Limitations and Future Work

We focused on creative story writing with short sessions (20--25 minutes); generalizability to other writing domains and longer sessions remains to be examined.

— Towards Cognitive Process-Aware Proactive Writing Support  (2608.30424 - Yoshida et al., 31 Aug 2026) in Section 6, Summary and Limitations

Whether repeated exposure to culturally misaligned AI drafts leads to durable changes in users' natural writing style, independent of AI assistance, remains unknown.

— AI-Generated Email Drafts Shift Culturally Distinctive Communication Styles in Professional Email  (2609.26403 - Sakai et al., 22 Sep 2026) in Section 5, Limitations and Future Work