<p>The Society of Automotive Engineers (SAE) defines Level 3 autonomous driving as conditional automation (CA), where the vehicle can perform a series of well-defined driving tasks under specific conditions but requires the driver to be ready to take over control at any time. However, the SAE definition is insufficiently user-centric, failing to clarify the specific forms that users expect from L3 autonomous vehicles. Since L3 autonomous driving remains an emerging field for users, eliciting its requirements through traditional methods such as user surveys and interviews alone is challenging. To address this, we proposed an innovative “expert-propose, user-choose” approach. Through this method, we identified and refined 11 key user-prioritized requirements, which can be categorized into four domains. Based on these findings, we further explored the user perspective and extracted three core requirements: “autonomous system safety,” “interface transparency,” and “driver attention.” Additionally, we defined the key characteristics of user-expected L3 autonomous vehicles: (1) the capability to safely and reliably perform autonomous driving tasks under specific conditions, (2) effective alerting mechanisms to ensure safe driver takeover, (3) maintenance of high situational awareness (SA) via transparent HMI, and (4) requirements for drivers to maintain sufficient attention and trust in the automated system. This study provides critical human-centered design references for automotive engineers developing L3 autonomous vehicles, advancing the further development and application of autonomous driving technology.</p>