由欧盟资助的斯洛伐克人工智能系统在2026年扫描了近100万巴西学童的脸部,引发了对隐私和伦理的担忧。 A Slovak AI system, funded by the EU, scanned nearly a million Brazilian schoolchildren’s faces in 2026, sparking privacy and ethics concerns.

2026年,一家斯洛伐克公司在欧盟资金支持下开发的面部识别系统在巴西巴拉那州1 700多所学校使用,扫描了近100万名儿童。 In 2026, a facial recognition system developed by a Slovak company backed by EU funds was used in over 1,700 schools across Brazil’s Paraná state, scanning nearly one million children. 尽管欧盟2024年的《大赦国际法》禁止在其境内进行这种监视,但该法缺乏出口限制,允许欧洲制造的大赦国际在没有监督的情况下部署到国外。 Despite the EU’s 2024 AI Act banning such surveillance within its borders, the law lacks export restrictions, allowing European-made AI to be deployed abroad without oversight. 该系统通过面部数据表示出勤率,但因不准确、与传统方法相比业绩不佳以及隐私风险(包括对获取社会福利的潜在影响)而面临批评。 The system, which marks attendance via facial data, has faced criticism for inaccuracies, poor performance compared to traditional methods, and privacy risks, including potential impacts on access to social benefits. 这一事件突显了全球AI治理中日益扩大的差距,并引起人们对在欧盟边界以外经营的欧洲技术公司的道德责任的关切。 The incident highlights a growing gap in global AI governance and raises concerns about the ethical responsibilities of European tech firms operating beyond EU borders.