每人配一个AI代理，实现直接民主的新可能
One Person, One Bot
- 为每位公民配备专属AI代理，代表其参与政治决策
- 提出可回溯的直接民主新范式，提升治理效率
- 适合关注未来社会治理与AI伦理的研究者
本文提出一种由代理型AI技术驱动的新型民主模式：为每位公民分配一个专属人工智能代理，作为其政治代表，以实现直接民主的回归。文章首先清晰阐述该模型的核心构想，随后探讨其与现有研究的关系、潜在挑战及可行性问题，并强调该理念的及时性与发展前景。论文主张应进一步推动该方向的研究与实践。
原文摘要 · Abstract (English)
This short paper puts forward a vision for a new democratic model enabled by the recent technological advances in agentic AI. It therefore opens with drawing a clear and concise picture of the model, and only later addresses related proposals and research directions, and concerns regarding feasibility and safety. It ends with a note on the timeliness of this idea and on optimism. The model proposed is that of assigning each citizen an AI Agent that would serve as their political delegate, enabling the return to direct democracy. The paper examines this models relation to existing research, its potential setbacks and feasibility and argues for its further development.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。