随着信息通信技术的普及与发展，无论是以刷脸支付为代表的商业应用场景，还是以刷脸出入校园为代表的公共应用场景，人脸识别技术已经在大学校园场景中得到广泛应用。在校大学生对人脸识别技术持何态度？不同场景中的风险感知与价值感知如何影响行为意愿？在资源有限的情况下，究竟是该设计更好的场景体验，还是做到更强的安全防护？

本研究首先通过文献研究法，基于行为理论、风险感知与价值感知、隐私权衡理论与场景/情境理论，对人脸识别技术的风险感知及其子维度隐私风险感知、系统风险感知进行概念界定；对人脸识别技术的价值感知及其子维度感知有用性、感知易用性进行概念界定。通过问卷调查法，利用问卷星平台线上收集了来自清华大学、北京大学、中国人民大学、北京师范大学、南京航空航天大学等高校在校大学生共512份有效样本，进行统计分析。

通过统计分析，验证了风险感知、价值感知与行为意愿、实际行为之间的关系。第一，在人脸识别不同校园场景中，风险感知与价值感知共同影响行为意愿，风险感知越低、价值感知越高，行为意愿越强烈。第二，在刷脸支付场景中，隐私风险感知与感知有用性共同影响行为意愿，即隐私风险感知越低、感知有用性越高，刷脸支付的行为意愿越强烈，同时，实际使用刷脸支付行为的频率也越高。第三，在刷脸出入校园场景中，隐私风险感知、感知有用性、感知易用性共同影响行为意愿，即隐私风险感知越低、感知有用性越高、感知易用性越高，刷脸出入校园的行为意愿越强烈；但是，行为意愿强烈与否都不影响实际使用刷脸出入校园行为的频率。这是由于在大多数情况下，学生对于是否使用刷脸出入校园功能不具有选择权，使用与否不能代表其是否愿意用。因此出现了不同于传统行为理论行为意愿正向影响实际行为的结果，行为意愿能否影响实际行为，需要视情境而定。

基于研究结果，反思行为理论、风险感知和价值感知、隐私权衡理论和场景/情境理论，可以得到四点结论。第一，不同的环境因素（校园环境，商业环境，社会环境等）会影响个体对于人脸识别技术的行为意愿；第二，当前“互联网时代无隐私”现象尤为突出，大学生隐私风险防范意识有待加强；第三，场景有用性是促进用户接受的关键，资源有限的情况下，提供更好的场景体验更能促进用户的使用意愿；第四，风险与价值的子维度因情境而变化，需要将更多风险和价值的维度纳入模型框架之中，以修正模型的适用范围。

基于结果讨论，从人脸识别技术各使用主体的角度，进一步给出人脸识别技术使用与监管策略的四条建议。首先，高校应当加强隐私风险的安全宣传教育，尽量减少人脸识别设备的供给。其次，企业应当积极承担社会责任，提供更多强有用性和弱隐私侵犯的场景。第三，高校作为弱监管者，需从源头进行监管，加强全过程数据安全保障。最后，政府作为强监管者，应当监管信息采集的必要性，从严惩治生物信息滥用等行为，提高违法成本。

With the popularization and development of information and communication technology, face recognition technology has been widely used in university campus scenes, whether in commercial application scenarios represented by face payment or in public application scenarios represented by face scan in and out of campus. What is the attitude of college students towards face recognition technology? How do perceived risk and perceived value affect behavior intention in different scenarios? In the case of limited resources, is it better to design a better scene experience or achieve stronger security protection?

First, through literature review, based on behavioral theory, perceived risk and perceived value, privacy trade-off theory and scenario/use context theory, the concept of variables are defined, including perceived risk of face recognition technology and its sub-dimensions of perceived privacy risk and perceived system risk, the perceived value of face recognition technology and its sub-dimension perceived usefulness and perceived ease of use. Through questionnaire survey, a total of 512 valid samples were collected from college students from Tsinghua University, Peking University, Renmin University of China, Beijing Normal University, Nanjing University of Aeronautics and Astronautics and other universities by using the questionnaire online platform, for statistical analysis.

Through statistical analysis, the relationship between perceived risk, perceived value, behavior intention and actual behavior is verified. First, in different campus scenes of face recognition, perceived risk and perceived value jointly affect behavior intention. The lower the perceived risk, the higher the perceived value, and the stronger the behavior intention. Second, in the face payment scenario, the perceived privacy risk and perceived usefulness together affect the behavior intention, that is, the lower the perceived privacy risk and the higher the perceived usefulness, the stronger the behavior intention of face payment, and at the same time, the higher the frequency of face payment. Third, in the case of face scan in and out of campus, the perceived privacy risk, perceived usefulness and perceived ease of use together affect the behavior intention, that is, the lower the perceived privacy risk, the higher the perceived usefulness and the higher the perceived ease of use, the stronger the behavior intention of face scan in and out of campus. However, the intensity of behavior intention did not affect the actual frequency of using face scanning to go to and from campus. This is because in most cases, students do not have the option to use the function, and whether they use it or not does not mean that they are willing to use it. Therefore, different from the traditional behavioral theory, the effect of behavior intention on actual behavior is positive. Whether behavior intention can affect actual behavior depends on the situation.

Based on the research results, four conclusions can be drawn from the reflection of behavior theory, perceived risk and perceived value, privacy trade-off theory, and scenario/use context theory. First, different environmental factors (campus environment, business environment, social environment, etc.) will affect individuals' behavioral willingness to face recognition technology. Second, the phenomenon of "no privacy in the Internet era" is particularly prominent, and college students' awareness of privacy risk prevention needs to be strengthened. Third, the usefulness of the scene is the key to promote user acceptance. In the case of limited resources, providing a better experience of the scene can better promote users' willingness to use it. Fourthly, as the sub-dimensions of risk and value change with the situation, more dimensions of risk and value need to be included in the model framework to modify the scope of application of the model.

Based on the results of the discussion, from the perspective of each user of face recognition technology, three suggestions on the use and supervision strategy of face recognition technology are further given. First of all, colleges and universities should strengthen security publicity and education on privacy risks and minimize the supply of facial recognition devices. Secondly, enterprises should actively undertake social responsibilities and provide more scenarios with strong usefulness and weak privacy violation. Third, as a weak regulator, colleges and universities need to conduct supervision from the source and strengthen data security guarantee throughout the process. Finally, as a strong regulator, the government should supervise the necessity of information collection, strictly punish the abuse of biological information and other behaviors, and raise the cost of illegal activities.