Value-Sensitive Delegation in Everyday AI Agent Use: Evidence from OpenClaw
Renkai Ma, Ruyuan Wan and colleagues at Cincinnati, Penn State, Arizona, South Carolina and Florida International analyze 73,093 first-person Reddit posts about using the OpenClaw agent to find which human values users care about when they delegate work.
Ask this paper
Method. Using Value Sensitive Design with LLM-assisted coding, each post is labeled with a human value, the agent aspect discussed, whether the value was fulfilled and the user outcome.
Six value groups. The 21 values form six groups, including Autonomous, Dependable and Affordable Operation, Bounded Reach, Reviewability and Equitable Access.
Where values cluster. Relative to each aspect's share of the corpus, values concentrate on the operating conditions users set around a run rather than on the agent's outputs.
Met versus unmet. Values were usually met when users described what the agent delivered, in five of six groups, and mostly unmet when users described supervising it, in all six.
Implication. The authors call this value-sensitive delegation and argue that agent evaluation should cover cost, access and oversight, not only task completion.
Abstract
Users increasingly delegate work to autonomous AI agents, yet evaluations typically measure task completion rather than the values users prioritize. Using Value Sensitive Design, we analyzed, with LLM assistance, 73,093 first-person Reddit posts about using OpenClaw, each for its human value, agent aspect, value fulfillment, and user outcome. The 21 values form six value groups, including Autonomous, Dependable, and Affordable Operation, Bounded Reach, Reviewability, and Equitable Access. Relative to each aspect's corpus share, values clustered not at the agent's outputs but at the operating conditions users set around a run. Values were usually met where users described what the agent delivered, in five of six groups, and mostly unmet where users described supervising it, in all six groups. We conceptualize this pattern as value-sensitive delegation. Supporting human values requires attention not only to what an agent accomplishes, but to the conditions users set around delegation, including cost, access, and oversight.