重庆社会科学

作为个性化新闻分发的底层支撑技术,算法推荐的科学技术特质决定了其并非中立,而是具有一定的意识形态属性。算法推荐的利用若不受监管,将会对主流意识形态传播、网络信息环境和用户的价值观念带来冲击,引发意识形态安全风险:平台流量至上的推送逻辑降低主流意识形态的引领力和凝聚力,以用户偏好为中心的内容推荐加剧了价值分化和观念隔离,过滤推荐引发价值迷失和信息操纵。有必要发挥法律的安全功能,通过强化网络平台的审核、监督义务和信息披露义务促进主流意识形态传播升级;通过赋予用户算法知情权、算法解释权打破信息茧房,建立多元信息环境;通过限制算法推荐的适用范围和算法审计监管构建"可信任"的算法,切实防范化解意识形态风险。

As the underlying technical support for personalized news distribution, the scientific and technological characteristics of algorithm recommendation determine that it is not neutral, but has certain ideological attributes.Algorithm recommendation has an impact on mainstream ideological dissemination, network information environment and user values, and contains ideological risks.The platform's communication mission of “flow first” reduces the leading power of mainstream ideology,preference recommendation intensifies concept isolation and value differentiation,filtering recommendations cause value loss and information manipulation.It is necessary to activate the safety function of law,promote mainstream ideology to lead dissemination by strengthening the audit, supervision and information disclosure obligations of the network platform,break the information cocoon by giving users the right to know the algorithm and the right to ask to explain the algorithm,by limiting the scope of application of algorithm recommendations and algorithm auditing to balance instrumental rationality and value rationality.