最新研究揭示,由人工智能生成的非自愿性深度伪造图像已不再局限于针对名人,而是日益严重地侵害着普通民众的权益 1。2024年7月对澳大利亚、英国和美国共计7230名成年人进行的调查显示,有7%的受访者表示曾遭受此类图像的侵害,另有3.2%的受访者承认自己创建、分享或以此威胁他人未经同意的AI亲密图像 1。受害者普遍遭受愤怒、羞耻和恐惧等心理创伤,且常因图像被社会视作“虚假”而遭到轻视,一名英国年轻女性直言:“没人关心、相信或倾听我” 1。
儿童群体受到的冲击尤为显著,数据显示截至2025年6月的18个月内,澳大利亚eSafety Commissioner收到的涉及数字篡改亲密图像的儿童报告数量,已超过此前七年总和的两倍 1。在这些儿童报告中,五分之四的案例涉及女孩 1。面对这一严峻形势,研究人员Nicola Henry和Gemma Stevens呼吁科技公司加强安全措施 1。政策层面也在积极回应,澳大利亚上月已将《数字注意义务》(digital duty of care)引入议会,旨在赋予监管机构打击“nudify”应用的权力 1。
A recent study indicates that non-consensual AI-generated deepfake images are increasingly harming ordinary individuals rather than solely targeting public figures. Researchers Nicola Henry and Gemma Stevens conducted a survey in July 2024 involving 7,230 adults across Australia, the United Kingdom, and the United States 1. The findings revealed that 7% of respondents had been victims of such content, while 3.2% admitted to creating, sharing, or threatening to share non-consensual AI intimate images 1.
The psychological impact on victims is severe, with many experiencing anger, shame, and fear, often compounded by social dismissal when their images are labeled as merely "fake" 1. One young woman from the UK expressed her isolation, stating, "nobody cared, believed me or listened" 1. The crisis extends significantly to minors, as the Australian eSafety Commissioner reported that in the 18 months leading up to June 2025, reports concerning digitally manipulated intimate images involving children exceeded twice the total number received over the previous seven years 1. Among these child-related reports, four out of five involved girls 1.
In response to these alarming trends, policymakers are considering stricter regulations. Last month, Australia introduced the Digital Duty of Care bill to parliament, which aims to empower regulators to combat "nudify" applications 1. The research team advocates for enhanced security measures by technology companies and supports legislative actions similar to those implemented in the UK to punish offenders effectively 1.
评论
还没有评论,欢迎留下第一条。