Instagram 的 AI 标注机制再度失灵

My Analysis of Instagram's AI Labeling Issues

Okay, here's what I've got. Instagram's whole point with these AI labels was to let people quickly identify synthetic content. Seems straightforward enough. But, judging from these reports, the system has completely gone haywire in the last few weeks. People are saying Meta's automatically slapping "AI Content" labels on images that weren't even created or edited using generative AI tools. Meanwhile, actual AI-generated stuff is slipping through the cracks, making everything look suspect on the platform. The reason for this tagging mess seems varied. A lot of users are seeing the label appear on images they’ve just lightly edited using tools like Canva's background remover – or even, really, on anything, or perhaps on images with slight changes. It's a significant issue. Let me go read the complete report at The Verge to understand it further.

Instagram 的可见 AI 标签原本旨在帮助人们一眼识别合成生成的内容。然而，在过去的几周里，用户接连反映该系统彻底失控了。他们表示，Meta 会自动为那些并非使用生成式 AI 工具创建或编辑的图片打上“AI 内容”（AI Content）标签。与此同时，真正的 AI 图像却成了漏网之鱼，让人感觉 Instagram 上的内容完全不可信了。被打上标签的具体原因似乎各不相同。许多人表示，使用 Canva 的“背景移除”（Background Remover）等工具编辑过，或是仅进行了微……的图片上也出现了该标签。请在 The Verge 阅读完整报道。