The field of human-AI collaboration is moving towards designing more effective and user-centered systems for knowledge work. Recent studies have highlighted the importance of understanding user needs and workflows to develop AI tools that can support complex tasks such as precedent search and document research. There is a growing recognition of the limitations of current AI systems, which often focus on end-to-end interaction and neglect the social context and iterative nature of knowledge work. As a result, researchers are exploring new approaches that prioritize accessibility, personalizability, and social awareness. Notably, the use of hybrid AI routers and multi-agent systems is being investigated for their potential to enhance efficiency, accuracy, and domain adaptability in applications such as machine translation and legal research.

Some noteworthy papers in this area include: The paper on designing human-AI systems for legal research, which presents an initial prototype for a precedent search tool that reflects the design implications derived from a qualitative study with legal practitioners. The paper on multi-agent systems for machine translation, which demonstrates the feasibility of multi-agent workflows in improving domain-adaptability and contextual awareness in translation tasks.