生成式人工智能的意识形态功能、风险及预前治理

The Ideological Functions,Risks,and Advance Governance of Generative Artificial Intelligence

生成式人工智能以其技术的底层性和通用性重塑意识形态实践的要素和结构,其专业化和小规模化后将成为广泛销售并使用的智能代理,将造就意识形态生产、传播、教育和动员新形态。由于专业权威弱化、算法复现数据性质谱系以及私有化的产权和应用造成了权威性与民主性、政治性与技术性、公共性与私有性矛盾,智能代理将在实践主体、模型构建和社会应用方面引起多重影响国家安全的意识形态风险。为治理风险,应立足未来提出预前策略,抛开为技术引发问题打补丁的思路。必须确立党管智能代理原则,强化主流意识形态权威性;构建可内嵌意识形态知识图谱,提升意识形态数据规范性;导入意识形态管理智能代理,实现社会应用全程风险处置。

Generative AI reshapes the elements and structure of ideological practice with its technological underpinnings and versatility,and in the future,after specialized and small-scaled,it will become an intelligent agent purchased and widely used by individuals,creating a new form of ideological production,dissemination,education and mobilisation.Due to the contradiction between authority and democracy,politics and technology,publicity and privacy caused by the weakening of professional authority,the spectrum of algorithmic reproduction of data nature,and privatised property rights and applications,intelligent agents will give rise to multiple ideological risks affecting national security in terms of the subject of practice,the construction of models,and social applications.In order to manage the risks,it is necessary to put forward preemptive strategies based on the future,leaving behind the idea of patching up the problems caused by technology.It is necessary to establish the principle of party-controlled intelligent agents to consolidate the authority of mainstream ideology;to introduce ideological knowledge mapping as a model development tool and social application specification;and to introduce ideological management intelligent agents as a central system of governance.