科技公司正在开发一种称为"追踪标记"（spymark）的隐形追踪技术，能够在图像、音频、文本和视频等数字媒体中嵌入人类无法感知的隐形信号。1 谷歌SynthID等系统可将这些信号编码成用户数据库标识符，用于追踪用户身份。1 与传统水印和EXIF等标准化元数据标签不同，追踪标记对用户完全不可见，且难以移除。1
这些隐形信号可以编码的个人信息范围广泛。谷歌SynthID-O的一个变体能在512x512像素的图像中编码136位有效载荷，其中包括64位数据库标识符。1 被编码的信息可涵盖用户记录、全名、IP地址、出生日期、物理地址和政治党派关联等数据。1 开源工具audiowmark自2018年以来便已能在音频文件中隐藏128位有效载荷。1
追踪标记技术并非全新发明。打印机追踪点是这类技术的早期例证，自1980年代以来就已被应用。1 然而，相关评论指出当前这一技术的广泛应用构成了隐私威胁，并呼吁将其识别为监控工具而非良性水印技术。1
A new generation of invisible tracking technology is emerging that goes far beyond traditional watermarking. Companies like Google have developed systems that embed imperceptible signals into images, audio, text, and video—a technique being termed "spymarks" rather than watermarks 1. These hidden markers can encode user database identifiers to track individual identity, fundamentally distinguishing them from visible or standard metadata approaches 1.
The technical capabilities of these systems are substantial. Google's SynthID-O variant can encode 136 bits of effective payload within 512x512 pixel images, including 64-bit database identifiers 1. Meanwhile, the open-source tool audiowmark has been capable of embedding 128-bit payloads in audio since 2018 1. Spymark signals can encode sensitive personal information such as user records, full names, IP addresses, birth dates, physical addresses, and political party affiliations 1.
What distinguishes spymarks from conventional tracking methods is their invisibility and resistance to removal. Unlike standardized metadata tags such as EXIF labels, these signals remain hidden from users and prove difficult to eliminate 1. Printer tracking dots represent an early precedent for this category of technology, dating back to the 1980s 1. Critics contend that such systems constitute surveillance tools rather than benign protective measures, raising significant privacy concerns about their deployment in digital media 1.
评论
还没有评论，欢迎留下第一条。