欧盟委员会更新 ERA 生成式 AI 负责任使用指南，新增智能体 AI 与评审规范

欧盟委员会发布了《ERA 关于在研究中负责任使用生成式人工智能的活页指南》更新版——这是一套建议，为研究界提供简单且可操作的指导，以促进生成式人工智能（AI）的采用和负责任使用。

The Commission has published an update of the ERA Living Guidelines on the responsible use of Generative AI in research - a set of recommendations offering simple and operational guidance to the research community to foster the adoption and responsible use of generative artificial intelligence (AI). 

这是该指南自第一版以来最全面的一次更新。活页指南经过修订，以应对生成式人工智能的快速发展，现纳入了关于代理式人工智能的指导，以及关于人工智能在同行评审和提案评估中应用的更详细指导。

This is the most comprehensive update since the first version of the guidelines. The Living Guidelines have been refined to address the rapid advancements in generative AI and now include guidance on agentic AI, as well as more detailed guidance regarding AI in peer review and proposal evaluation. 

借助代理式人工智能，工具正变得更加自主，能够处理日益复杂的任务，这使得监督、问责和可靠性问题愈发重要。在利用人工智能评估资助提案方面，此前的指导已根据科学界的需要得到澄清和扩展。

With agentic AI, tools are becoming more autonomous and capable of handling increasingly complex tasks, making questions of oversight, accountability and reliability all the more important. In the context of using AI to evaluate funding proposals, the previous guidance has been clarified and expanded in response to the needs of the scientific community.

指南的简单实用方法得以保持，并通过修订后的结构得到强化。研究诚信原则，如问责、透明和负责，仍是主要参照，并被转化为应对生成式人工智能使用的实用指导。

The simple and practical approach of the Guidelines is maintained and reinforced with a revised structure. Research integrity principles, such as accountability, transparency and responsibility, are still the main reference and are translated into practical guidance to address the use of generative AI.

背景

Background

该指南于2024年首次发布。此后经过多次修订，上一次修订是在2026年5月。

The Guidelines were first published in 2024. Since then, they have undergone several revisions, the previous one in May 2026. 

这项工作是在欧洲研究区（ERA）各国和各利益相关方的投入与协作下制定的。其包容性方法仍是该过程的核心，并辅以一个开放的反馈流程，以便持续为指南提供贡献和建议。

This work is developed with the input and collaboration of countries and stakeholders across the European Research Area (ERA). Its inclusive approach remains central to the process, supported by an open feedback process for ongoing contributions and suggestions to the Guidelines.

更多信息

More information 

指南

Guidelines

情况说明书

Factsheet

科学中的人工智能

AI in science

注意转载授权 · 阅读原文。中文正文为 AI 翻译，请以原文为准。

注意转载授权 · 阅读原文