Neural-hash-collider

Security and privacy

adversarial hash collision tool

SHOW HN2021

623points
351comments
#31of 2,533 in 2021
Top 2%of 2021 by points

How HN took it

RoastedHalf or more of the comments are critical

8 positive, 47 neutral or asking, 1 mixed, 61 critical, out of 117 comments about the project. Thread score -0.38.

Idea: 3 for, 29 againstPrivacy: 1 for, 22 againstExecution: 3 for, 9 againstDesign: 0 for, 1 against

The thread, comment by comment

Each row is a top-level comment and its replies, in HN's order. Select a cell to open that comment on HN.

positive neutral or mixed critical about another commenter or something else the maker

What HN said

What people liked

  • Commenters praise the work as a demonstration that pressures Apple to change the system.

    From 2 comments
    • Well done! Hopefully all of this progress toward demonstrating how easy it is to manipulate neural hash will get Apple to rollback the update...
      SCUSKU on HN
    • Great work, I hope people keep hacking the system to lower the system's credibility. This idea is just beyond insane, and the plan to have manual check on user's photos on their…
      ipiz0618 on HN

Objections

  • Perceptual hashes are not collision resistant, so a second hash layer will not fix the design.

    From 2 comments
    • I don't see how this is fixable on their end. Several people have suggested simply layering several different perceptual hash systems, with the assumption that it's difficult to…
      blintz on HN
    • I've seen a lot of comments "muddying the waters" (intentionally or not) about whether hash colliders like the one demonstrated above can be used to carry out an attack. So I…
      saithound on HN
  • Hash collisions could be used to frame people or flood reviewers with false positives.

    From 3 comments
    • The integrity of this entire system now relies on the security of the CSAM hash database, which has just dramatically increased in value to potential attackers. All it would take…
      vermilingua on HN
    • In addition to the attacks, such as converting legit image to be detected as CSAM (false positive) or circumventing detection of the real CSAM image (false negative), which have…
      xucheng on HN
    • Some people seem to be confused why a hash collision of a cat and a dog matters. Here's a potential attack: share (legal) NSFW pictures that are engineered to have a hash…
      throwaway384950 on HN

Requests

  • Apple should roll back or change the NeuralHash scanning update.

    From 1 comment
    • Well done! Hopefully all of this progress toward demonstrating how easy it is to manipulate neural hash will get Apple to rollback the update...
      SCUSKU on HN

Questions

  • Whether a secret second hash function would stop collision attacks.

    From 2 comments
    • I am strongly against Apple’s decision to do on-device CSAM detection, but: wasn’t there a secondary hash whose database is not shared? In theory you need to collide with both to…
      jchw on HN
    • I've seen a lot of comments "muddying the waters" (intentionally or not) about whether hash colliders like the one demonstrated above can be used to carry out an attack. So I…
      saithound on HN

Alternatives named

  • Apple's NeuralHash CSAM detection system

    From 1 comment
    • Ongoing related threads: Apple defends anti-child abuse imagery tech after claims of ‘hash collisions’ - https://news.ycombinator.com/item?id=28225706 - Aug 2021 (401 comments)…
      dang on HN

Category, keywords, summaries and labels on this page were made by AI (Thread Reader v2) and not checked by hand. Each point links to the comments it came from.

Add a badge to your READMEShown on HackerShow
[![Shown on HackerShow](https://hackershow.org/badge.svg)](https://hackershow.org/show/28229291)