欧盟2025年的《AI法》禁止社会评分,并规定了高风险AI的透明度、人的监督以及数据安全,全球遵守,最高罚款7%。 The EU’s 2025 AI Act bans social scoring and mandates transparency, human oversight, and data security for high-risk AI, with global compliance and up to 7% fines.

2025年,欧盟的《AI法案》充分执行了高风险AI系统的规则,禁止社会评分等做法,并要求透明度、人力监督和严格的数据治理。 In 2025, the EU’s AI Act fully enforces rules for high-risk AI systems, banning practices like social scoring and requiring transparency, human oversight, and strict data governance. 为了遵守,公司必须确保问责制和减轻风险,处罚最高为全球收入的7%。 To comply, companies must ensure accountability and risk mitigation, with penalties up to 7% of global revenue. 该法案适用于全球,正在塑造国际人工智能政策. The Act, applicable globally, is shaping international AI policy. 作为回应,IONOS等欧洲供应商推出了主权的AI平台,如IONOS AI 模型枢纽和INONOS GPT——德国开发的安全、开放源码、符合GDPR的AI工具,避免外国数据访问,使企业和公共机构拥有数据主权。 In response, European providers like IONOS launched sovereign AI platforms—such as the IONOS AI Model Hub and IONOS GPT—offering secure, open-source, GDPR-compliant AI tools developed in Germany, avoiding foreign data access and enabling data sovereignty for businesses and public institutions.