Connected-vehicle and roadside telemetry enable low-latency safety navigation, and traffic-optimisation services at the edge, but finegrained mobility streams (locations, speeds, events, and contexts) create high re-identification and linkage risk when accessed by multiple stakeholder domains. We present a Zero Touch Network and Service Management (ZSM) integrated, policy-driven data-collection service that operationalises mobility data governance through intent-based automation. Stakeholders submit high-level collection intents (purpose, fields, spatial/temporal granularity, latency, and utility targets); a policy engine evaluates and rewrites intents into compliant, effective intents; and a plan generator compiles them into executable data-collection pipelines deployable within a ZSM closed loop. Experiments on Beijing taxi mobility traces execute 87,500 DP-protected releases and achieve 1.26\% relative error for Road Safety Authority (RSA) at ϵ = 8.0, while DP-Stochastic Gradient Descent (DP-SGD) risk scoring reaches 0.97 ± 0.03 test accuracy at ϵ = 0.5,