论网络暴力的平台技术治理

On the Plaform Technology Governance of Cyberbullying

网络暴力的行为特殊性导致既有法律治理难见成效。无论是传统个案司法救济模式,还是网信部门“信息内容”监管模式,均无法有效预防、控制网络暴力。平台技术治理成为目前的主要方案并被相关立法所规定。技术治理可实现对网络暴力的事前预防及信息准确识别;平台作为“守门人”是控制网络暴力信息的最佳着力点,其数据垄断地位也使得相关技术方案有落地可能性。在技术治理方式上,立法规定了网络暴力识别模型、用户保障机制、信息阻断等。其中,建立网络暴力识别模型是关键,其包括对内容的识别模型和对高危账号的识别模型。此外,可探索性地引入社交机器人技术,生成良性信息来对抗网络暴力信息。同时不可忽视网络暴力治理技术所带来的风险。针对“不良信息型”网络暴力的认定,应将个体感受、人物关系、上下文情境等要素纳入识别模型;针对平台用户权利风险,引入“技术正当程序”与比例原则,保障用户的被通知权、知情权、算法解释权等权利并限制平台权力;针对社交机器人安全风险,要考虑网络暴力治理场景的正当性,加强其价值观引导,确保内容可靠并控制传播边界;此外,要合理界定注意义务主体范围,由大型平台与超大型平台承担网络暴力治理义务。

The specificity of cyberbullying poses challenges to the effectiveness of existing legal governance frame-works.Both the traditional individual judicial relief model and the regulatory model of"information content"by the internet information office are insufficient in effectively preventing and controlling cyberbullying.Currently,platform technological governance has emerged as a primary solution mandated by relevant legislation.Technological governance enables early warning of cyberbullying and the accurate identification of information.Platforms,acting as"gatekeepers",play a crucial role in controlling violent information on the internet due to their monopolistic position in data,making technological solu-tions against cyberbullying feasible.Legislative requirements for technological governance solutions include establishing models for identifying cyberbullying,implementing user protection mechanisms,and enabling information blocking.Among these requirements,developing an effective model for identifying different types of online violence is paramount;this in-cludes models for both content analysis and identification of high-risk accounts.Additionally,exploring the use of social robots can generate positive counteractive information against abusive content.However,it is important not to overlook po-tential risks associated with technology-based measures against online violence.To identify"armful information"cyberbul-lying,it is imperative to incorporate elements such as individual perceptions,interpersonal relationships,and contextual factors into recognition models.Addressing the risks to user rights on platforms requires"technological due process"and the principle of proportionality to safeguard users'rights to notification,informed consent,algorithmic transparency,and to limit plaform power.To mitigate security risks associated with social bots,it is essential to consider the legitimacy of cy-berbullying governance scenarios,reinforce value guidance,ensure content reliability,and control the boundaries of infor-mation dissemination.Furthermore,it is crucial to clarify the scope of duty bearers,with the responsibility for cyberbully-ing governance being shouldered by large and super-large digital platforms.