英国政府计划引入由德国公司Cognitec开发的人工智能面部年龄识别技术来筛选入境移民，但多家人权组织对该方案提出严厉批评。[1] 根据英国内政部数据，2025年5月有755名儿童在抵达英国时被错误认定为成人。[1] 政府部门承认即使是最先进的系统也可能存在长达30个月的误差幅度。[1]
人权组织指控这些AI系统对撒哈拉以南非洲儿童存在明显的种族偏见。[1] 一项调查表明Cognitec的系统倾向于高估未成年人的年龄。[1] 被错误判定为成人的儿童面临被纳入成人移民系统的风险，可能遭遇拘留和遣返。[1] 权利倡导组织Humans for Rights Network的Maddie Harris指出这是一种"廉价流程"，不会改善现状。[1] 研究人员Petra Molnar将面部年龄识别技术与19世纪的伪科学"颅相学"进行对比。[1] 难民和移民儿童联盟指出，创伤可能会改变儿童的外观，进一步降低这类技术的可靠性。[1]
The UK government's plan to deploy facial age recognition technology developed by German firm Cognitec to screen migrants has drawn sharp criticism from human rights groups over accuracy concerns. As of May 2025, the Home Office recorded that 755 children were incorrectly identified as adults upon arrival in Britain [1], highlighting the real-world consequences of such screening failures. Though government officials acknowledge that even the most sophisticated systems may contain a margin of error of up to 30 months [1], rights advocates argue this level of inaccuracy makes the technology unsuitable for consequential immigration decisions.
Independent testing has revealed that the Cognitec system exhibits racial bias against Sub-Saharan African individuals, tending to overestimate the age of minors [1]. Maddie Harris of the Humans for Rights Network characterizes the approach as a "cheap process" that will not resolve the underlying problem [1]. Critics have drawn comparisons between facial age estimation and discredited 19th-century pseudoscience such as phrenology [1], questioning the scientific validity of automated age determination. Additionally, organizations working with refugee and migrant children note that trauma can visibly alter a child's appearance [1], suggesting that biographical factors beyond chronological age may confound algorithmic assessment.