Europe’s AI labeling and transparency rules are now in effect
Europe’s AI labeling and transparency rules are now in effect
欧洲人工智能标签与透明度规则正式生效
The European Union has ushered in some additional rules that aim to make it easier for people to identify chatbots and AI deepfakes online. The new transparency obligations under the bloc’s landmark AI Act came into effect on August 2nd, requiring companies to disclose when people are interacting with AI models, and if content has been generated or altered by them.
欧盟已出台一系列补充规则，旨在让人们更容易识别在线聊天机器人和人工智能深度伪造内容。根据欧盟具有里程碑意义的《人工智能法案》（AI Act），新的透明度义务已于 8 月 2 日生效，要求企业在用户与人工智能模型交互时，以及内容由人工智能生成或篡改时，必须进行披露。
The transparency rules differ between providers (companies that develop and market AI systems) and deployers (platforms and services that use those AI systems), though some companies like Meta and SpaceXAI are classified as both.
透明度规则在“提供者”（开发和销售人工智能系统的公司）与“部署者”（使用这些人工智能系统的平台和服务）之间有所不同，尽管像 Meta 和 SpaceXAI 这样的公司被同时归类为这两者。
Providers must design AI systems to explicitly notify users when they’re dealing with AI instead of a human, “unless this is obvious,” and include machine-readable marks on synthetic audio, image, video, and text so that it can be detected as artificially generated or manipulated. Likewise, deployers are required to label any AI-generated or manipulated image, audio, and video deepfake content that’s designed to look real.
提供者必须设计人工智能系统，在用户与人工智能而非人类交互时明确通知用户（“除非显而易见”），并在合成的音频、图像、视频和文本中包含机器可读的标记，以便将其识别为人工生成或篡改的内容。同样，部署者必须对任何旨在看起来真实的、由人工智能生成或篡改的图像、音频和视频深度伪造内容进行标注。
“The rapid development of generative and interactive AI systems is making it increasingly difficult to distinguish AI interactions and AI-generated content from human-created and authentic content,” the European Commission said in its transparency guidelines. “People should know when they are interacting with AI or exposed to AI-generated content. This will help them make informed decisions, calibrate their trust and reliance on AI and avoid misinformation or deception.”
欧盟委员会在其透明度指南中表示：“生成式和交互式人工智能系统的快速发展，使得区分人工智能交互与人工智能生成内容，以及人类创作与真实内容变得越来越困难。人们应该知道自己何时在与人工智能交互，或接触到人工智能生成的内容。这将有助于他们做出明智的决定，校准对人工智能的信任和依赖，并避免错误信息或欺骗。”
The Commission has also developed a set of AI disclosure labels that tech platforms can adopt in accordance with the AI Act rules, alongside examples of when they should be used. These mirror similar labels that have already been introduced by TikTok, Instagram, and Facebook. The EU-made icons have been introduced to prevent platforms from having to design their own, and while the EU examples are optional to use, the Commission pointedly notes that its labeling requirements “are not.”
欧盟委员会还制定了一套人工智能披露标签，科技平台可以根据《人工智能法案》的规则采用这些标签，并附带了使用场景示例。这些标签与 TikTok、Instagram 和 Facebook 已经引入的类似标签相呼应。欧盟制作这些图标是为了避免平台自行设计，虽然欧盟提供的示例是可选的，但委员会明确指出，其标签要求是“强制性的”。
Companies that fail to comply with these new transparency rules risk facing fines of up to €15 million (about $17.2 million), or 3 percent of their global annual turnover. While the new transparency rules are enforceable immediately for new AI systems, any models and services that launched before August 2nd have a four-month grace period until December 2nd to comply.
未能遵守这些新透明度规则的公司将面临最高 1500 万欧元（约合 1720 万美元）或其全球年度营业额 3% 的罚款。虽然新的透明度规则对新的人工智能系统立即生效，但任何在 8 月 2 日之前发布模型和服务的公司，将拥有四个月的宽限期，需在 12 月 2 日前完成合规。