首页|基于深度合成技术的文化科技伦理风险识别与善治机制

基于深度合成技术的文化科技伦理风险识别与善治机制

扫码查看
生成式人工智能技术的应用使人工智能直接参与数字文化的生产环节,既成为新质生产力,也增加了文化科技伦理风险.生成式人工智能在内容生成上导致的"幻觉"叠加深度合成在内容传播过程中的虚假效应,进一步扩展了人工智能在数字文化领域的潜在风险;即在核心内容生产层面涉及新型文化意识形态安全风险、文化内容安全风险、文化版权安全风险,在外围传播层面涉及文化数据安全风险和新型文化信息安全风险.生成式人工智能带来的文化科技伦理风险根源于智能时代文化资本的数字化、人机关系倒置化和认知方式的变化.文化科技伦理风险的善治机制亟须在全球治理体系框架下,整合核心技术研发、预见性制度规范与多元主体协同联动,形成政府、企业、公共三方的"组合拳"效应,以人机协同、善治向上的治理机制回应生成式人工智能时代数字文化新质生产力变革.
Risks Recognition and Its Governance Mechanism on Cultural Ethics Embodied in Deep Synthesis Technology
The application of generative artificial intelligence generated content technology allows AI to directly participate in the production of digital culture,becoming a new form of productive force while also generating risks on cultural ethics embodied in technology.AIGC-induced"Hallucinations"in content generation,combined with the false effects of deep synthesis during content dissemination,further expand the potential risks of AI in the digital culture domain.These risks include new types of cultural ideological security risks,cultural content security risks,and cultural copyright security risks at the core content production level,as well as cultural data security risks and new types of cultural information security risks at the peripheral dissemination level.Fundamentally,risks on cultural ethics embodied in technology brought by AIGC stem from the digitization of cultural capital,the inversion of human-machine relationships,and changes in cognitive modes in the intelligent era.To address these ethical risks,it is imperative to develop a governance mechanism that integrates core technology research and development,anticipatory institutional norms,and coordinated multi-stakeholder collaboration within the framework of global governance.This"combined punch"approach involving the government,enterprises,and the public aims to respond to the transformative productive forces of digital culture in the AIGC era through a governance mechanism that promotes human-machine collaboration and sound governance.

artificial intelligence generated contentdeep synthesisethics risksgovernance mechanisms

解学芳、陈思函

展开 >

同济大学 人文学院,上海 200082

生成式人工智能 深度合成 伦理风险 善治机制

国家社会科学基金重大项目

23&ZD087

2024

南昌大学学报(人文社会科学版)
南昌大学

南昌大学学报(人文社会科学版)

CSSCICHSSCD北大核心
影响因子:0.413
ISSN:1006-0448
年,卷(期):2024.55(4)