澳大利亚一家科技公司推出了名为SAFE的机器学习系统，在全国1000多家早期教育服务机构中使用，通过分析教育工作者的日常观察记录来识别可能处于危险中的儿童[1]。该系统掌握了约200,000名托儿行业工作者的行为记录[1]，在儿童安全面临重大挑战的背景下应运而生。
澳大利亚近期发生多起托儿工作者虐童案件，引发了行业范围的安全改革。2024年2月27日，澳大利亚儿童安全法律改革生效[1]。昆士兰州工作者Ashley Paul Griffith因对73名儿童的307项罪名在2024年被判终身监禁[1]；悉尼托儿工作者Hamish Alexander Edward Tait上月被逮捕，面临320多项指控，涉及150名儿童[1]。为应对这一局面，政府拨款2.26亿澳元用于全国CCTV试点计划[1]。
然而SAFE系统的推出引发了关于儿童隐私、数据安全和算法偏见的重大争议[1]。澳大利亚信息专员办公室将在12月10日前制定《儿童在线隐私守则》[1]。该系统的首席执行官Ruby O'Rourke本人是童年虐待幸存者，曾在2019年参选联邦议会[1]。
An Australian technology company has deployed a machine learning system called SAFE across more than 1,000 early childhood education services nationwide, designed to identify children who may be in danger by analyzing daily observations recorded by educators [1]. The system holds behavioral records on approximately 200,000 childcare industry workers [1].
The initiative emerges amid significant safety reforms in the sector. Australia implemented major childcare safety law reforms on February 27, 2024 [1], and the government has allocated A$226 million for a nationwide CCTV pilot program [1]. These measures follow high-profile incidents including the 2024 lifetime sentence imposed on Queensland worker Ashley Paul Griffith for 307 offenses involving 73 children [1], and the recent arrest of Sydney childcare worker Hamish Alexander Edward Tait, who faces more than 320 charges involving 150 children [1].
The SAFE system's launch has sparked considerable concern regarding child privacy, data security, and algorithmic bias [1]. The Australian Information Commissioner's Office is developing a Children's Online Privacy Code, with a deadline of December 10 [1]. The company's CEO, Ruby O'Rourke, is a childhood abuse survivor who previously ran for federal parliament in 2019 [1].