I noticed a bit of panic around here lately and as I have had to continuously fight against pedos for the past year, I have developed tools to help me detect and prevent this content.

As luck would have it, we recently published one of our anti-csam checker tool as a python library that anyone can use. So I thought I could use this to help lemmy admins feel a bit more safe.

The tool can either go through all your images via your object storage and delete all CSAM, or it canrun continuously and scan and delete all new images as well. Suggested option is to run it using --all once, and then run it as a daemon and leave it running.

Better options would be to be able to retrieve exact images uploaded via lemmy/pict-rs api but we’re not there quite yet.

Let me know if you have any issue or improvements.

EDIT: Just to clarify, you should run this on your desktop PC with a GPU, not on your lemmy server!

  • snowe@programming.dev
    link
    fedilink
    English
    arrow-up
    5
    ·
    1 year ago

    You may not have meant it, but you strongly implied something of the sort. But since this is not what you’re suggesting I’m curious to hear what your optimal approach to those problem would be here.

    Optimal approach is to use the existing systems that are used by massive corporations to solve this problem already. I know everyone on lemmy hates that, but this isn’t something to mess around with. The reason this is optimal is because NCMEC provides the hashes only to these companies. You’re not going to be able to get the hashes (this is a good thing… imagine some child abuser getting access to these hashes and then using them to evade detection). So if you can’t get these hashes (and you shouldn’t want them either) then you should use a service that has them. It is by far the best way to filter and has been proven time and time again to be successful.

    The easiest is CloudFlare’s, and yes, you will have to use them as your DNS which I also understand a vast majority of admins hate. But there are other options as well

    • PhotoDNA
    • Safer
    • Facebook PDQ

    Because access to the original hash databases is considered sensitive, NCMEC will not provide these to smaller platforms. Neither will Microsoft provide the source code of its PhotoDNA algorithm except to its most trusted partners, because if the algorithm became widely known, it is thought that this might enable abusers to bypass it.

    In that article, it actually points out that a solution called Safer that uses machine learning and image recognition has very flawed results and is incredibly biased. So if these massive platforms can’t get this kind of image recognition right then it’s probably best to not waste money and time on it. The article even points out that for smaller platforms it’s not worth it.

    We also know in general terms that machine learning algorithms for image recognition tend to be both flawed overall, and biased against minorities specifically. In October 2020, it was reported that Facebook’s nudity-detection AI reported a picture of onions for takedown. It may be that for largest platforms, AI algorithms can assist human moderators to triage likely-infringing images. But they should never be relied upon without human review, and for smaller platforms they are likely to be more trouble than they are worth

        • db0@lemmy.dbzer0.comOP
          link
          fedilink
          English
          arrow-up
          2
          ·
          edit-2
          1 year ago

          Have you already registered for this services and using them on your lemmy? If so the success is something displayed in time.

          • snowe@programming.dev
            link
            fedilink
            English
            arrow-up
            0
            ·
            1 year ago

            They have been used on millions of websites already. It’s pretty clear that it works. It doesn’t need to be used on lemmy to prove it works. And my application is currently in review so no I haven’t used it. But that really doesn’t matter. Especially if you’re comparing it to a tool written by one person that has been out for a few days.

            • db0@lemmy.dbzer0.comOP
              link
              fedilink
              English
              arrow-up
              1
              ·
              1 year ago

              The situation Has dramatically changed in the the past year. I am telling you but you seem to be in denial. Likewise currently you’re unprotected. As such my previous statement applies: good luck!

              • snowe@programming.dev
                link
                fedilink
                English
                arrow-up
                0
                ·
                1 year ago

                The situation Has dramatically changed in the the past year. I am telling you but you seem to be in denial.

                What situation? You haven’t told me anything about anything changing in the last year.

                Likewise currently you’re unprotected. As such my previous statement applies: good luck!

                Huh? What previous statement? You’re not protected. You just think you are! You literally claimed that your product has lots of false positives. It most definitely has false negatives too.

                  • snowe@programming.dev
                    link
                    fedilink
                    English
                    arrow-up
                    0
                    ·
                    1 year ago

                    …. That’s not what CSAM is. It’s literally in the name. abuse. CSAM requires abuse of a child. Just like pictures of your kids in the bath isn’t illegal, AI generated content is going to need to be provably linked to abuse in order to meet the definition. I seriously doubt it ever will be. You seem to have a fundamental misunderstanding of the entire situation, overreacted so much so that you might have gotten people in serious legal trouble, and are now trying to justify it because you think that the current solutions aren’t going to protect against CSAM.

                    I literally linked a software that does what yours does, but is already used, already proven to work, and also has the benefit of being the backup to the hash based mechanism instead of being the primary mode of filtering.

                    Dude, I was trying to be nice how I approached you earlier, but you clearly are unwilling to accept that you have made a massive mistake here, not only by not reading up on the laws, but not even fundamentally understanding the problem space. This is not good software engineering.