首页 > AI前沿 > Scaling public input on AI policy with pluralistic agents

Scaling public input on AI policy with pluralistic agents

arXiv自然语言 2026-11-12 05:35 19 阅读 查看原文
用户需要我将给定的纯文本新闻正文排版成语义化的HTML片段。规则如下: 1. 只输出HTML片段本身,不要Markdown围栏,也不要解释。 2. 每个自然段落用`

`包裹,长文用`

`/`

`划分小节。 3. 列表内容用`
  • `,引用用`
    `,代码用`
    `,重要句子可用``。
    4. 保留原文所有信息,不得增删改内容。
    5. 不要输出``/``/``等文档级标签。
    
    让我仔细分析给定的文本,识别其中的自然段落和需要强调的重要句子。
    
    文本:
    "AI is being deployed across a vast range of application domains, cultural environments, and human stakeholders. Safe and responsible AI development requires continuously accounting for diverse perspectives. We develop and test pluralistic agents, AI agent representatives that deliberate on behalf of individuals to align AI systems with diverse human perspectives. Pluralistic agents could help people contribute to many more discussions than they could directly participate in, on technical topics that would require training, and on policy choices that rely on private information that corporations and governments cannot readily share. Our experiments identify opportunities, risks, and perceptions of pluralistic agents. Agents were able to reliably shift policies in the direction of their represented persona's preferences, but the extent of each agent's influence was sensitive to the instructions from its persona, such as having more influence when instructed to be more verbose or less cooperative. This suggests that influence in AI-assisted deliberation could vary based on how participants prompt their agents and how the system is designed. In a preregistered study, 746 humans participated in the pluralistic agents process. Participants viewed pluralistic agents as useful and fair, preferring agents over the baseline of providing input to a single AI reasoning model. Most participants preferred participating themselves over either pluralistic agents or the baseline, but pluralistic agents recovered 68% of the baseline-direct gap in reported satisfaction and 82% in fairness. Because pluralistic agents can be used in many real-world contexts where direct discussion is not, they could be a promising tool to scale AI safety policy development. We encourage cautious exploration of this method, playing to its strengths as an amplifier of human input rather than a replacement."
    
    分析内容结构:
    这是一段连续的文本,似乎是一个摘要或引言。它提到了:
    1. AI部署和安全、负责任的发展需要多样化的观点。
    2. 开发和测试“pluralistic agents”(多元智能体)。
    3. 多元智能体的好处。
    4. 实验结果:识别机会、风险和看法。
    5. 智能体能够可靠地转移政策,但影响力取决于指令。
    6. 这表明AI辅助审议中的影响力可能有所不同。
    7. 一项预注册的研究,746名人类参与。
    8. 参与者认为多元智能体有用且公平,比基线更好。
    9. 大多数参与者更喜欢自己参与,但多元智能体在满意度和公平性方面恢复了很大的差距。
    10. 多元智能体可以在直接讨论不可行的真实世界情境中使用,是扩展AI安全政策开发的有前景的工具。
    11. 鼓励谨慎探索这种方法,作为人类输入的放大器而非替代品。
    
    这看起来是一篇学术论文或报告的摘要。我需要将其排版成语义化的HTML。
    
    由于它是一段连续的文本,没有明显的标题或小节,我可以将其视为一个整体段落,或者根据逻辑划分小节。但规则说“长文用`

    `/`

    `划分小节”,这里并没有现成的小标题。我可以添加小节标题吗?规则说“保留原文所有信息,不得增删改内容”。添加小节标题会引入原文没有的内容,所以可能不应该添加。但规则也说“长文用`

    `/`

    `划分小节”,这意味着如果原文有小节,就使用它们。原文没有小标题,所以我可能不需要添加。 不过,我可以将这段文本分成几个逻辑段落。原文是一个大段落,但为了可读性,我可以根据主题将其拆分成几个自然段落,并保持内容不变。 让我看看原文的结构,它实际上是一整段。但为了排版,我需要用`

    `包裹每个自然段落。原文只有一个自然段落,所以我可以将其整体放入一个`

    `中。 但规则说“长文用`

    `/`

    `划分小节”。原文没有小节标题,所以我不应该添加。我可以只使用`

    `。 让我考虑是否可以将原文拆分成几个`

    `,基于逻辑停顿。例如: - 第一句到“diverse human perspectives.”可以是一个段落。 - 然后关于pluralistic agents的好处。 - 然后关于实验。 - 然后关于研究结果。 - 最后关于结论。 但拆分段落需要谨慎,不能改变内容。我可以将原文分成几个`

    `,每个`

    `包含一个或多个句子,但保持句子顺序和内容不变。 让我尝试按逻辑拆分: 1. "AI is being deployed across a vast range of application domains, cultural environments, and human stakeholders. Safe and responsible AI development requires continuously accounting for diverse perspectives." 2. "We develop and test pluralistic agents, AI agent representatives that deliberate on behalf of individuals to align AI systems with diverse human perspectives." 3. "Pluralistic agents could help people contribute to many more discussions than they could directly participate in, on technical topics that would require training, and on policy choices that rely on private information that corporations and governments cannot readily share." 4. "Our experiments identify opportunities, risks, and perceptions of pluralistic agents." 5. "Agents were able to reliably shift policies in the direction of their represented persona's preferences, but the extent of each agent's influence was sensitive to the instructions from its persona, such as having more influence when instructed to be more verbose or less cooperative." 6. "This suggests that influence in AI-assisted deliberation could vary based on how participants prompt their agents and how the system is designed." 7. "In a preregistered study, 746 humans participated in the pluralistic agents process." 8. "Participants viewed pluralistic agents as useful and fair, preferring agents over the baseline of providing input to a single AI reasoning model." 9. "Most participants preferred participating themselves over either pluralistic agents or the baseline, but pluralistic agents recovered 68% of the baseline-direct gap in reported satisfaction and 82% in fairness." 10. "Because pluralistic agents can be used in many real-world contexts where direct discussion is not, they could be a promising tool to scale AI safety policy development." 11. "We encourage cautious exploration of this method, playing to its strengths as an amplifier of human input rather than a replacement." 这样拆分成多个段落是合理的,因为每个段落都表达一个相对独立的意思,而且没有改变内容。但原文是一整段,我这样拆分是否算“增删改内容”?严格来说,只是将段落拆分,内容没有增删改。所以是允许的。 现在,关于重要句子,我可以用``来强调一些关键句子。哪些是重要的?例如: - "We develop and test pluralistic agents" 可能是重要的。 - "Agents were able to reliably shift policies..." 可能是重要的。 - "pluralistic agents recovered 68% of the baseline-direct gap..." 可能是重要的。 - "they could be a promising tool to scale AI safety policy development" 可能是重要的。 - "playing to its strengths as an amplifier of human input rather than a replacement" 可能是重要的。 但规则说“重要句子可用``”,这意味着我可以选择一些关键句子进行强调。我会选择几个核心论点。 现在,构建HTML。我不需要`

    `/`

    `,因为原文没有小节标题。但规则说“长文用`

    `/`

    `划分小节”,如果原文没有,我就不使用。所以只使用`

    `和``。 让我写出HTML片段:

    AI is being deployed across a vast range of application domains, cultural environments, and human stakeholders. Safe and responsible AI development requires continuously accounting for diverse perspectives.

    We develop and test pluralistic agents, AI agent representatives that deliberate on behalf of individuals to align AI systems with diverse human perspectives. Pluralistic agents could help people contribute to many more discussions than they could directly participate in, on technical topics that would require training, and on policy choices that rely on private information that corporations and governments cannot readily share.

    Our experiments identify opportunities, risks, and perceptions of pluralistic agents. Agents were able to reliably shift policies in the direction of their represented persona's preferences, but the extent of each agent'