Skip to content
Trending storyWatching

Image trust and provenance verification in the generative AI era

1 report1 sourceupdated 2 days ago

What happened

AI digest

On November 21 a Hacker News front-page post argued that generative AI makes image authenticity harder to judge. The author calls for default skepticism, with creators supplying provenance credentials to rebuild trust. Drawing on forensic work on a missile strike video, the post covers C2PA signatures, Google SynthID watermarks and sensor signing on some new iPhones as provenance tools. It also floats per-frame signature checks for live video calls, pairing provenance with tamper forensics. The post does not say this has been implemented.

Written by AI from the coverage · updated 2 days ago

Coverage

Follow the reports to see the story from different sides.

Nov 21
  1. Hacker News front page
    I Don't Trust Visual Recordings Anymore. But One Day I Will Again

    生成式 AI 让影像真实性更难判断,作者主张转向默认怀疑,并由创作者提供来源凭证来建立信任。作者结合导弹袭击视频的取证经历,解释 C2PA 签名记录、Google SynthID 水印及部分新款 iPhone 的传感器签名如何支持来源核验。实时视频通话也可探索逐帧签名验证,让来源认证与检测篡改痕迹的取证分析相互配合。

Heat over time

Heat now 3·Comparable peak 9(Nov 21 12:00)·Comparable change over 24 hours -50%

02.557.510Nov 2112:00Nov 2201:00Nov 2213:00Nov 2302:00

The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.