Traditional knowledge elicitation techniques (e.g., interviews) are a crucial element of modern research, but are often accompanied by several shortcomings or biases (e.g., memory misattribution) which makes choosing the right technique imperative. Reciprocal knowledge elicitation describes the beneficial application of mutual knowledge elicitation and knowledge provision between members of a human-agent team. The design of reciprocal knowledge elicitation remains so far unexplored. This workshop aims to identify and scope the relevant context for knowledge elicitation techniques in general and to elevate techniques by applying human-agent collaboration. For the main outcomes, we aim for a domain-specific overview of relevant attributes and a formalized representation of knowledge elicitation techniques.

Call For Paper

While knowledge elicitation techniques are rarely directly addressed by prevailing research questions, they play a crucial and irreplaceable role in Hybrid Human-Artificial Intelligence research. We want to spotlight knowledge elicitation techniques by inviting researchers and practitioners to submit:

• Experience Reports of former work describing how the choice of their selected knowledge elicitation technique led to success (or even endangered the quality of results by introducing biases). If possible, experience reports shall also address by which measures the specific knowledge elicitation technique was chosen and evaluated. Preferably, the reports appropriately address the human-agent collaboration or communication model in general.
• Vision Papers. As (reciprocal) knowledge elicitation is still a young concept but steadily growing and accelerated by Agentic AI, we also accept vision papers that describe how such techniques can be designed, implemented, measured, or evaluated.

In alignment with the CEUR guidelines, experience reports and vision papers will be accepted in the short paper format of a length between 5 to 9 pages. A LaTeX template is hosted on Overleaf. The reviewing process is single-blinded. Authors should not anonymize their submission. Also, we encourage to add links to supplementary material.