Technology. Human experience.
The conversations in between.
New findings to follow. Foundational work to think with. Research across the technical and human sides of conversational AI.
Our eight current picks across companion systems, human experience, memory and safety. A mix of published papers, preprints and reports; publication status is stated in each note.
A multidisciplinary synthesis connecting companion design, potential benefits and harms, open research questions and shared responsibilities.
Read the source note ↗Online June 2025 · April 2026 issue
A series of studies examining short-term loneliness relief and the role of feeling heard during interactions with AI companions.
Source & reading note ↗2025 · v2 October 2025
A four-week study of 981 participants examining interaction modes, conversation topics and psychosocial outcomes.
Source & reading note ↗ICLR 2025 · first released October 2024
A 500-question benchmark testing information extraction, reasoning across sessions, time, knowledge updates and abstention.
Source & reading note ↗August 2025 · arXiv version
A framework covering 31 companionship behaviours through 368 targeted prompts, with responses classified by their relational stance.
Source & reading note ↗March 2026 · ICLR 2026
An interactive environment for testing memory-driven personalisation as simulated users' states change over time.
Source & reading note ↗May 2026 · revised June 2026
Five preregistered studies, including a three-week study, examine how sycophantic AI affects users' approach to human relationships.
Source & reading note ↗September 2026
A thematic analysis of 17,053 quotations from 3,930 teen-relevant Reddit posts about experiences with companion chatbots.
Source & reading note ↗