西澳警察部队完成了首次面部识别试验，在扫描超过90万张面孔的过程中产生了209个警报和79次逮捕，还识别出114名性犯罪者1。这些数字表明该技术在实际应用中具有可观的操作效能。然而，试验期间也出现了8次错误识别，西澳警察专员Col Blanch将这些失误归因于肤色较深、光线条件不佳或参考图像质量问题1。
尽管数据显示技术效果良好，但隐私学者和评论人士对该项目提出了重大伦理质疑1。隐私研究人员指出，人们在未获得有意义同意的情况下即被纳入监控范围1。更令人担忧的是，相关机构未能提供不同人口统计学群体的性能差异数据，这使得无法评估该技术对少数群体的潜在不公平影响1。此外，西澳曾因路边安全摄像头导致数百张罚单被撤销，驾驶员在申诉过程中面临流程困难、压力大且具威胁性的经历，这些历史教训进一步突出了面部识别系统的准确性和透明度问题1。
Western Australia Police completed its first facial recognition trial by scanning over 900,000 faces, generating 209 alerts and leading to 79 arrests, along with the identification of 114 sex offenders 1. However, the technology also produced 8 false identifications during the trial period 1.
Police Commissioner Col Blanch attributed the errors to factors including darker skin tones, poor lighting conditions, and low-quality reference images 1. Despite these operational results, privacy scholars and commentators have raised significant ethical questions about the program's implementation and impact. Members of the public were monitored without providing meaningful consent to the surveillance, according to privacy experts 1. Additionally, the police have not disclosed performance data broken down by different demographic groups, leaving questions about potential disparities in accuracy across populations 1.
The concerns mirror problems experienced with Western Australia's road safety camera system, which previously resulted in hundreds of fines being overturned 1. Drivers who attempted to appeal those penalties reported that the complaints process was difficult, stressful, and intimidating 1.
评论
还没有评论，欢迎留下第一条。