三项研究探讨自主 AI 如何重塑责任、专业判断与现实决策

继续使用 Google、Facebook 或 Apple,即表示您同意使用条款。
三项研究探讨自主 AI 如何重塑责任、专业知识与现实世界决策
由 Sadie Harley 编辑,Andrew Zinin 审阅
本文已根据 Science X 的编辑流程和政策进行审阅。编辑在确保内容可信度的同时,强调了以下属性:
添加到首选来源
谁应为自主人工系统的行为负责,其判断应被赋予多少权威,当它在物理世界中行动时又会发生什么变化?关于人工智能的公众与科学辩论正日益聚焦于智能体式 AI(或 Agentic AI),即能够追求目标并自主决定采取哪些行动来实现这些目标、而无需每一步都需要人类干预的系统。
从进行在线购物的数字助手,到管理客户请求或协调工作活动的系统,这些技术的快速发展开辟了新的可能性,同时也引发了日益紧迫的伦理和社会问题。
米兰理工大学(Politecnico di Milano)的跨部门项目 META——面向科学与技术的社会科学与人文——研究了智能体式 AI 的发展和使用对人类与人工系统之间关系构成的挑战。
智能体式 AI 引发的主要伦理问题之一,是责任归属问题,这也是 Giacomo Zanotti 研究的焦点。例如,设想一个人工助理为了不超出预算,自主决定取消一项已经规划好的行程,而用户并未明确要求它这样做。
Fabio Iapaolo、Giacomo Zanotti 和 Camilla Quaresmini。图片来源:Politecnico di Milano
在这种情况下,谁该负责?这项研究揭示了一个根本性的张力:这些系统越自主,要确定谁应当为其行为负责就越复杂。
智能体式 AI 也在改变我们生产和评估知识的方式,由此带来的新问题正是 Camilla Quaresmini 研究的焦点。例如在医院场景中,一个 AI 系统可能会分析医疗报告和检验结果,并提出临床评估意见,而这个意见可能被赋予比医生的观察或患者的陈述更高的权威。
该研究识别出 AI 在知识过程中的三种不同角色,区分了组织信息的系统、产出评估的系统,以及主动参与研究活动的系统。研究还指出,保持人类控制并不保证人的评估会比人工系统的评估获得更高的权威。
最后,当智能体式 AI 不再仅仅在数字环境中运作,而是直接与物理世界(包括我们的城市)互动时,又会带来什么变化?这正是 Fabio Iapaolo 关于自动驾驶汽车研究的核心问题,自动驾驶也是所谓“物理智能体式 AI”最具代表性的例子之一。
这项研究挑战了决策可以归因于单一实体(无论是人还是人工系统)的观点,转而表明决策如何通过传感器、算法、控制设备与周围环境之间的互动而形成。
这一视角凸显了拓宽 AI 伦理反思的必要性:不仅要考虑单个决策及其直接后果,还要考虑这些系统借以在世界中行动的技术、物质和社会关系。
“在智能体式 AI 系统正被应用于从医疗保健到出行和工作组织等日益多样化领域的背景下,这三项研究展示了人工系统日益增长的自主性如何提出远超任务与决策自动化的问题,”Zanotti、Quaresmini 和 Iapaolo 解释道。
从这些系统所执行行动的责任,到赋予其判断的权威,再到决策通过技术与物理环境之间的互动而形成的方式,该研究项目提供了关键工具,用于理解正在发生的转型并应对其伦理和社会影响。
这篇报道出自谁之手?
生命科学与生态学理学学士。拥有微生物学实验室背景,并具备石油、天然气和可再生能源行业的医药新闻经验。完整简介 →
物理学硕士,拥有研究经验。长期的科学新闻爱好者。在 Science X 编辑工作的成功中发挥关键作用。完整简介 →
引用:三项研究探讨自主 AI 如何重塑责任、专业能力与现实世界决策(2026年11月19日),检索于2026年11月19日,来源:https://techxplore.com/news/2026-11-autonomous-ai-reshapes-responsibility-expertise.html
本文件受版权保护。除为私人学习或研究目的进行的任何合理使用外,未经书面许可,不得复制任何部分。内容仅供参考。
研究发现,人们认为智能AI助手应对结果负责
研究发现,智能体式 AI 考验数据保护法的极限
如何让“智慧城市”技术合乎伦理地运作
AI与社会的共同演化:研究探讨机遇与风险
研究表明,AI的广泛采用掩盖了更深层次的问题
莲茎和生物膜利用蒸发的水产生电力,为小型电子设备供电
迄今建造的最快软体机器人无需电子器件即可运作
羽量级飞行机器人自带传感器和计算机
微观“聚合物纤维森林”有助于为电子设备散热
结构电池电极可在弯曲的无人机支腿内实现储能
迷迭香与魔方:气味如何帮助我们在混合现实中记住所学内容
获取即时摘要文本(要点)
AI 自主性增强使其行为的责任归属变得复杂。人类监督并不能确保人类判断对 AI 评估保有权威。在自动驾驶汽车中,决策源于技术组件与环境之间的交互,这要求伦理分析超越个体行动者和即时结果。
本摘要由 LLM 自动生成。完整免责声明
译文由英文原文机器翻译,术语按行业惯例统一,可在文末展开原文核对。
查看英文原文(Three studies examine how autonomous AI reshapes responsibility, expertise and real-world decisions)
By continuing with Google, Facebook or Apple, you agree to the Terms of Use.
Three studies examine how autonomous AI reshapes responsibility, expertise and real-world decisions
edited by Sadie Harley, reviewed by Andrew Zinin
This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility:
Add to Preferred Sources
Credit: Image generated by the editorial team using AI for illustrative purposes.
Who is responsible for the actions of an autonomous artificial system, how much authority should be attributed to its judgments, and what changes when it acts in the physical world? The public and scientific debate on artificial intelligence is increasingly focusing on agentic AI (or Agentic AI), namely systems capable of pursuing goals and autonomously deciding which actions to take to achieve them, without requiring human intervention at every step.
From digital assistants that make online purchases to systems that manage customer requests or coordinate work activities, the rapid development of these technologies opens up new possibilities, while raising increasingly urgent ethical and social questions.
The interdepartmental project META—Social sciences and humanities for science and technology at Politecnico di Milano has studied the challenges that the development and use of agentic AI pose to the relationship between human beings and artificial systems.
Among the main ethical questions raised by agentic AI is that of the attribution of responsibility, which was the focus of research by Giacomo Zanotti. Consider, for example, an artificial assistant that, in order to stay within a budget, autonomously decides to cancel an already planned trip, without the user having expressly asked it to do so.
Fabio Iapaolo, Giacomo Zanotti and Camilla Quaresmini. Credit: Politecnico di Milano
In this case, who would be responsible? The research highlighted a fundamental tension: the more autonomous these systems become, the more complex it is to establish who should be held responsible for their actions.
Agentic AI is also transforming the way we produce and assess knowledge, raising new questions that were the focus of research by Camilla Quaresmini. In a hospital setting, for example, an AI system might analyze medical reports and test results and suggest a clinical assessment, which could be attributed greater authority than a doctor's observations or the patient's account.
The research identified three different roles of AI in knowledge processes, distinguishing between systems that organize information, produce assessments or actively participate in research activities. It also highlighted how maintaining human control does not guarantee that people's assessments will be granted greater authority than those of artificial systems.
And finally, what changes when agentic AI does not merely operate in digital environments, but interacts directly with the physical world, including our cities? This is the question at the heart of research by Fabio Iapaolo on self-driving cars, one of the most emblematic examples of so-called "physical agentic AI."
The research challenged the idea that decisions can be attributed to a single entity, human or artificial, showing instead how they take shape through interactions among sensors, algorithms, control devices and the surrounding environment.
This perspective highlights the need to broaden ethical reflection on AI, considering not only individual decisions and their immediate consequences, but also the technical, material and social relationships through which these systems act in the world.
"In a context in which agentic AI systems are being applied in increasingly diverse fields, from health care to mobility and the organization of work, the three studies show how the growing autonomy of artificial systems raises questions that go far beyond the automation of tasks and decisions," explain Zanotti, Quaresmini and Iapaolo.
From responsibility for the actions performed by these systems to the authority attributed to their judgments, and the way decisions take shape through the interaction between technologies and the physical environment, the research project offers critical tools for understanding the transformations underway and addressing their ethical and social implications.
Who's behind this story?
BSc Life Sciences & Ecology. Microbiology lab background with pharmaceutical news experience in oil, gas, and renewable industries. Full profile →
Master's in physics with research experience. Long-time science news enthusiast. Plays key role in Science X's editorial success. Full profile →
Citation: Three studies examine how autonomous AI reshapes responsibility, expertise and real-world decisions (2026, November 19) retrieved 19 November 2026 from https://techxplore.com/news/2026-11-autonomous-ai-reshapes-responsibility-expertise.html
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.
People hold smart AI assistants responsible for outcomes, study finds
Agentic AI tests the limits of data protection law, study finds
How to make 'smart city' technologies behave ethically
Coevolution of AI and society: Study explores opportunities and risks
Widespread AI adoption masks deeper problem, study shows
Lotus stems and biofilms generate electricity from evaporating water to power small electronics
The fastest soft robot ever built doesn't need electronics to function
Featherweight flying robot carries its own sensors and computers
Microscopic 'polymer fiber forest' helps keep electronics cool
Structural battery electrode enables energy storage inside curved drone legs
Rosemary and Rubik's Cubes: How a scent helps us remember what we learn in mixed reality
Get Instant Summarized Text (Gist)
Greater AI autonomy complicates attribution of responsibility for its actions. Human oversight does not ensure that human judgments retain authority over AI assessments. In self-driving cars, decisions emerge from interactions among technical components and environments, requiring ethical analysis beyond individual actors and immediate outcomes.
This summary was automatically generated using LLM. Full disclaimer
This site uses cookies to assist with navigation, analyse your use of our services, collect data for ads personalisation and provide content from third parties. By using our site, you acknowledge that you have read and understand our Privacy Policy and Terms of Use.
原文为英文,正文由自动抓取并保留原文。
原文发布于 TechXplore(Three studies examine how autonomous AI reshapes responsibility, expertise and real-world decisions),标题、摘要与正文由机器翻译,术语按行业惯例统一,未经人工逐句校订。
