首页|人工智能驱动的虚假信息:现在与未来

人工智能驱动的虚假信息:现在与未来

扫码查看
新的人工智能功能正在迅速增加虚假信息的数量、速度、多样性和病毒性,使内容创建过程和虚假信息活动的实施自动化.随着这些功能不断改进和扩散,它们破坏了对可验证事实的信任,进一步威胁到民主治理,并导致公民怀疑公共生活中真相的可能性.总体来说,人工智能通过制造深度伪造品、大型语言模型的开发与部署、人工智能增强的社交机器人和微定向等方式生产、重构和传播虚假信息.而虚假信息的问题无法采取单一的解决方案,无论是心理上还是技术上,有效应对这一挑战需要了解技术、媒介和人类行为的融合因素.我们可通过多维度安排实现问责分配、重视互联网平台的把关人作用、保障新闻媒体的正常信息供应、用优质信息淹没虚假信息、提高公众的网络素养等途径和方法,探索虚假信息的可能治理路径.
AI-Driven Disinformation:Present and Future
New capabilities in artificial intelligence are rapidly increasing the quantity,speed,diversity,and virality of disinformation,automating the processes of content creation and disinforma-tion campaigns.As these capabilities continue to improve and proliferate,they erode trust in verifia-ble facts,further threatening democratic governance and fostering citizen skepticism about the possi-bility of truth in public life.Overall,artificial intelligence contributes to the production,reconstruc-tion,and dissemination of disinformation through methods such as deepfakes,the development and deployment of large language models,AI-enhanced social bots,and microtargeting.Addressing the challenge of disinformation requires a multifaceted approach,considering the converging factors of technology,media,and human behavior both psychologically and technically.Possible governance pathways for disinformation can be explored through strategies such as multidimensional accountabili-ty arrangements,emphasizing the role of gatekeepers on internet platforms,ensuring the normal sup-ply of information by news media,drowning out disinformation with high-quality information,and enhancing the digital literacy of the public.

disinformationartificial intelligencedeepfakelarge language modelsocial botmicrotargeting

胡泳

展开 >

北京大学新闻与传播学院 北京 100871

虚假信息 人工智能 深度伪造 大型语言模型 微定向

2024

南京社会科学
南京市社会科学界联合会 南京市社会科学院 中共南京市委党校

南京社会科学

CSTPCDCSSCICHSSCD北大核心
影响因子:0.998
ISSN:1001-8263
年,卷(期):2024.(1)
  • 1
  • 49