AI解码口罩人脸易出错，误判风险高
AI-Powered Facial Mask Removal Is Not Suitable For Identification
- 用商用AI模型去除人脸口罩，生成清晰图像
- 生成人脸与真实身份匹配率不足50%
- 适合关注AI安全与身份识别风险的研究者
近期，网络群体调查中使用生成式AI增强低质量视觉证据。一起高关注度案件中，社交媒体用户传播了联邦特工在致命射击事件中的“AI去口罩”图像，引发广泛误认。针对此类事件，我们开展大规模分析，评估商业AI面部去口罩技术的有效性与风险，重点考察生成面部是否可被可靠匹配至真实身份。
原文摘要 · Abstract (English)
Recently, crowd-sourced online criminal investigations have used generative-AI to enhance low-quality visual evidence. In one high-profile case, social-media users circulated an "AI-unmasked" image of a federal agent involved in a fatal shooting, fueling a wide-spread misidentification. In response to this and similar incidents, we conducted a large-scale analysis evaluating the efficacy and risks of commercial AI-powered facial unmasking, specifically assessing whether the resulting faces can be reliably matched to true identities.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。