自动驾驶汽车能够精确记录用户大量的公开活动数据，使得对隐私的挑战在所难免，引发了人们对其数据隐私的担忧。当前该领域下个人数据的有效保护与合理利用之间已然形成了一种难以协调的冲突关系。然而，我国现行法律规范多偏向原则性和纲领性的规定，很多具体的实际操作问题尚未涉及，且在实践层面还存在着明显的不适用，个人数据保护还留待不少问题亟需解决。总结起来，即个人数据识别边界模糊、“知情—同意”原则落实困难，以及权利、义务和责任规范不够完备等。本文关切的核心问题是自动驾驶汽车新技术的发展给个人数据保护带来的挑战是否能够被现有的法律规定充分调整和控制，是否有必要以及如何制定新的具体规定。鉴于自动驾驶场景汇集着来自数据安全、数据利用效率以及社会公共利益等多种价值选择，对个人数据保护基本规范要求的落实还须首先从理念上加以调整，进而实现合理的制度安排。为此，本文引入场景一致性理论，主张基于高度的场景化，以综合、动态的视角实现“场景公正”；并在此基础上，根据总结出的实践困境，针对性探索更为精细的“场景化治理”路径。具体而言，首先应对自动驾驶汽车个人数据的概念进行重新审视，将“人”与“车”之间的连接距离作为判断是否属于个人数据的主要标准；同时，由于匿名化技术仍存在局限性，匿名化标准尚不明晰，应当引入“合理可能”标准，并辅之以具象化的参考指标。针对“知情—同意”原则难以落地的困境，应当区分个人数据的敏感程度和使用场景，以分层同意和动态同意为中心，同时要求数据控制者履行持续有效的信息披露义务。最后，应当细化数据主体的权利条款，在未来的相关立法中进一步明确权利的使用范围和行使条件，以解决个人数据保护“承诺太多，兑现太少”的困境。对于数据控制者义务规范、监督机制方面控制不足的难题，可以从培育数据控制者的内部治理机制和构筑有效的外部监督两个层面出发，将个人数据保护要求嵌入到整体的数据安全防范体系之中，实现激励相容的制度效果。
Autonomous vehicles can accurately record a large amount of public activity data of users, which makes the challenge to privacy inevitable and causes people to worry about their data privacy. Currently, the effective protection and rational use of personal data in this field have formed an irreconcilable and conflicting relationship. However, the current legal norms are mostly principle-oriented and programmatic provisions, and many specific practical operational issues have not yet been addressed. There is still obvious inapplicability in practice. To summarize, the protection of personal data still leaves a lot of problems that need to be solved urgently, namely, the blurred boundaries of personal data identification, the difficulty of implementing the principle of “informed consent”, and the incomplete provisions of rights, obligations and responsibilities. The core issue of this paper is whether the challenges to personal data protection brought by the development of new autonomous driving technologies can be adequately adjusted and controlled by existing legal provisions, whether and how new specific provisions are necessary.Given that the autonomous driving context brings together multiple value choices from data security, data utilization efficiency, and social public interests, the implementation of basic requirements for personal data protection needs to be adjusted from the conceptual perspective, and then realize a reasonable institutional arrangement. Therefore, this paper introduces contextual integrity theory, advocating for a highly contextualized approach to achieve “contextual integrity” from a comprehensive and dynamic perspective; and on this basis, based on the practical difficulties summarized, exploring a more refined path of “contextual based governance”.Specifically, first of all, the concept of personal data of autonomous vehicles should be re-examined, and the connection distance between the “person” and the “car” should be taken as the main criterion to judge whether it belongs to personal data or not; at the same time, due to the limitations of anonymization technology, the standard of is still unclear, and the criterion of “all the means reasonably likely to be used” should be introduced and supplemented with concrete reference indicators. In response to the difficulty of implementing the principle of “informed consent”, it is necessary to distinguish the sensitivity of personal data and the context of its use, with tiered consent and dynamic consent as the center, while requiring data controllers to fulfill the obligation of continuous and effective information disclosure. Finally, the provisions of data rights should be refined, and the scope of use and conditions for exercising rights should be further clarified in future relevant legislation, in order to solve the dilemma of “too many promises, too little fulfillment” in personal data protection. For the problem of insufficient control in terms of obligations and supervision mechanisms for data controllers, personal data protection requirements can be embedded into the overall data security prevention system from two aspects: cultivating internal governance mechanisms of data controllers and building effective external supervision, to achieve incentive compatible institutional effects.
