Memoro is a wearable AI memory assistant that continuously captures audio through a bone conduction headset, transforming everyday conversations into a structured, living memory. Users can recall details through natural language voice queries, receive proactive reminders extracted from their own speech, and review end-of-day summaries of their conversations, all without manual note-taking or screen interaction.
The system was developed across two research phases. In the first (CHI 2024), we introduced Memoro's core architecture: an LLM-powered retrieval agent that enables both explicit voice queries and a novel Queryless Mode, where the system infers what the user needs from conversational context alone. A study with 20 participants showed that Memoro increased recall confidence and reduced task load while preserving conversational quality, with an 85% reduction in response length compared to a standard LLM baseline.
In the second phase, we extended the system into Memoro V2, adding proactive commitment extraction and hierarchical daily summaries, features informed by a need-finding survey of 100 older adults. We then deployed Memoro V2 in a 10-day field study with ten participants aged 60 to 86, who wore the system freely in their everyday lives. Over 140 hours of recorded speech and 460+ memory interactions revealed sustained engagement and natural daily rhythms: planning in the morning, support during the day, and review at night. Participants used the system to manage appointments, recover forgotten details, prepare for sensitive conversations, and reflect on their own communication patterns. The study surfaced a central insight: when always-on AI delivers immediate, recurrent value, older adults' perception of continuous recording shifts from intrusive to integral.
Honorable Mention at Proceedings of the CHI Conference on Human Factors in Computing Systems (CHI '24). Association for Computing Machinery, New York, NY, USA, Article 450, 1–18. https://doi.org/10.1145/3613904.3642450
Honorable Mention in Student Category by FastCompany World Changing Ideas '24