数字西部世界?斯坦福 AI 智能体小镇在 Apache-2.0 协议下正式开源
今年早些时候,斯坦福和谷歌的研究人员以《模拟人生》游戏为灵感,创建了一个 AI 智能体小镇;目前该 AI 小镇已在 Apache-2.0 协议下正式开源。
研究人员在模拟城镇中添加了 25 个生成式智能体 (Generative Agents),这 25 个角色由 ChatGPT 和自定义代码控制,以高度逼真的行为独立地生活。在 ChatGPT 的支持下,每个人都有自己独特的身份、记忆和行为,并且可以独立交互,但他们都不会意识到自己是生活在模拟中。
生成式智能体起床,做早餐,然后去上班;艺术家作画,作家写作;他们形成意见、互相关注并发起对话;他们在计划第二天时会记住并反思过去的日子…… 在评估中,这些生成式智能体会产生可信的个体行为和突发性社会行为:例如,从用户指定的一个 Agent 想要举办情人节派对的单一概念开始,Agents 会在接下来的两个时间里自主地传播派对邀请几天、结识新朋友、互相约出参加聚会的日期,并协调正确的时间一起出现在聚会上。
根据介绍,为了实现这一目标,研究人员严重依赖一个用于社交互动的 LLM,特别是 ChatGPT API。此外,他们还创建了一种架构,用记忆和经验来模拟思维,然后让 Agents 在世界中自由互动。人类也可以与它们互动。“用户可以观察并干预 Agents 计划他们的日程、分享新闻、建立关系和协调小组活动。”
作为研究的一部分,该小组还聘请了人类评估人员观看模拟重播,以衡量人工智能代理根据其环境和经验做出可信行为的程度,包括 “可信的计划、反应和想法” 以及 “信息扩散、关系形成和 Agent 在社区不同区域的协调”。
研究人员还要求人类以角色扮演的方式,模仿他们观看到的 Agent 的声音来回答采访问题;结果发现,”完整的 Generative Agents 架构” 产生的结果比参与者角色扮演的效果更可信。
这也就引出了一些技术的伦理影响和风险问题。研究人员警告称,存在形成不适当的 “准社会关系 (parasocial relationship)”、错误推论的影响、加剧与生成人工智能相关的现有风险,以及在设计过程中过度依赖 Generative Agents 的风险等风险。
论文地址:《Generative Agents: Interactive Simulacra of Human Behavior》
Believable proxies of human behavior can empower interactive applications ranging from immersive environments to rehearsal spaces for interpersonal communication to prototyping tools. In this paper, we introduce generative agents–computational software agents that simulate believable human behavior. Generative agents wake up, cook breakfast, and head to work; artists paint, while authors write; they form opinions, notice each other, and initiate conversations; they remember and reflect on days past as they plan the next day. To enable generative agents, we describe an architecture that extends a large language model to store a complete record of the agent’s experiences using natural language, synthesize those memories over time into higher-level reflections, and retrieve them dynamically to plan behavior. We instantiate generative agents to populate an interactive sandbox environment inspired by The Sims, where end users can interact with a small town of twenty five agents using natural language. In an evaluation, these generative agents produce believable individual and emergent social behaviors: for example, starting with only a single user-specified notion that one agent wants to throw a Valentine’s Day party, the agents autonomously spread invitations to the party over the next two days, make new acquaintances, ask each other out on dates to the party, and coordinate to show up for the party together at the right time. We demonstrate through ablation that the components of our agent architecture–observation, planning, and reflection–each contribute critically to the believability of agent behavior. By fusing large language models with computational, interactive agents, this work introduces architectural and interaction patterns for enabling believable simulations of human behavior.

