Hello everyone,

We unfortunately have to close the !lemmyshitpost community for the time being. We have been fighting the CSAM (Child Sexual Assault Material) posts all day but there is nothing we can do because they will just post from another instance since we changed our registration policy.

We keep working on a solution, we have a few things in the works but that won’t help us now.

Thank you for your understanding and apologies to our users, moderators and admins of other instances who had to deal with this.

Edit: @[email protected] the moderator of the affected community made a post apologizing for what happened. But this could not be stopped even with 10 moderators. And if it wasn’t his community it would have been another one. And it is clear this could happen on any instance.

But we will not give up. We are lucky to have a very dedicated team and we can hopefully make an announcement about what’s next very soon.

Edit 2: removed that bit about the moderator tools. That came out a bit harsher than how we meant it. It’s been a long day and having to deal with this kind of stuff got some of us a bit salty to say the least. Remember we also had to deal with people posting scat not too long ago so this isn’t the first time we felt helpless. Anyway, I hope we can announce something more positive soon.

  • 𝕯𝖎𝖕𝖘𝖍𝖎𝖙@lemmy.world
    312·
    1 year ago

    They aren’t asking devs to be admins or for admins to be devs. They specifically called out the developers because code exists to filter child sexual abuse material, disseminated by organizations such as the FBI and law enforcement, which can be implemented for image uploading.

    Yeah? I doubt this is true but I could be wrong. You make it sound like preventing CSAM is as simple as importing a library, something I find dubious. Companies have been trying to filter out this material in an automated fashion for decades and yet they still have to employ humans to do it manually because automated means don’t really work. This is why companies like Reddit, facebook have trust and safety teams to do this work.

    Edit: I goggled and could not find this database. I’m thinking it’s a myth.

    NOBODY in this comment section is advocating for uploading fucking child sexual abuse material. That is a strawman you are setting up. Nobody is advocating for allowing the uploading of child sexual abuse material, or for the “material to be up on lemmy instances”. NOBODY is suggesting that a single instance going down is “the world is ending”. NOBODY is asking for “100’s of mods to specifically address this one user’s posting of CSAM”.

    ahem there were users who uploaded CSAM. Those are the users who were advocating for uploading CSAM, becuase they uploaded CSAM.

    I’m literally arguing with people who are saying that they shouldn’t have shut down the community because it’s big and that shutting down the community (not CSAM) poses an threat to the fediverse. Maybe, but CSAM poses a legal threat, which is much greater than the threat of low engagement.

    You’re setting up a strawman argument nobody is proposing. The criticism is that, at this moment, the developers of Lemmy have not implemented a method for automatically vetting uploaded images for CSAM without requiring “100’s of mods”, which is what resulted in the condition that “taking the community down is the only option here”.

    Yeah, that doesn’t exist, as I’ve mentioned previously. You make it sound like getting CSAM off lemmy was as simple as writing some code - if it were, why doesn’t facebook and reddit do this?

    Perhaps the wording of the original post was not precise and accurate enough for your full and complete understanding of the intent and meaning behind it. In this post, I have attempted to elucidate that intent and meaning to a degree which I hope is understandable to you.

    You’re not understanding how CSAM detection works or is handled.

    The grim reality is this: cameras exist, children exist, adults exist, the internet exists, and the second that a crime is committed, it is not added to an FBI database. If such as FBI database existed and IF it was useful (and not just a database of hashes for bit-perfect copies of CSAM) and IF it were updated when evidence of the crime surfaces… IF all of those things are true, THEN it means there’s still likely a huge swath of CSAM material still out there, that could be posted at any time, and that would NOT be detected.

    Again ask yourself, IF such a database existed, then WHY does reddit, twitter, facebook, hell, why doesn’t every or any site use it?

    Pedophiles, instead of downvoting me, why not explain yourself?

    • wanderingmagus@lemm.ee
      6·
      1 year ago

      As another commenter posted below:

      But tools do exist. PhotoDNA by Microsoft. Although much more user-friendly implementation if you use Cloudflare, related links:

      As far as I am aware, every major site does use it in addition to manual vetting for any flagged “borderline” or “uncertain” results caught up in the filter.

      • 𝕯𝖎𝖕𝖘𝖍𝖎𝖙@lemmy.world
        18·
        1 year ago

        As far as I am aware, every major site does use it in addition to manual vetting for any flagged “borderline” or “uncertain” results caught up in the filter.

        I think this is where you could be wrong here. I appreciate the links, I’ll look into those in more detail. My best understanding is that these tools generate so many false-positives and false-negatives that it’s not worth using them. It may be a first line of defense until real humans get to see them, but my point is that humans are still needed. When humans are included because the system isn’t 100%, it means humans do the labor and as such, with limited time, humans need to determine when they can do the labor - sometimes shutting down a community is the best way to stop the flood while they clean up the mess.

        • Richard@lemmy.worldEnglish
          3·
          1 year ago

          This is just a matter of confirmation bias from your side now. You stubbornly refuse to accept factual information very helpfully delivered to you by users who have many better things to do than respond to your inquiries, and you dogmatically refuse to acknowledge that there are advanced and reliable automation tools available for the use case in question. And while you do all that, you belittle the other users in the community by referring to your supposedly superior knowledge and experience, however somehow failing to provide any data or secondary sources to back up your claims.

          • 𝕯𝖎𝖕𝖘𝖍𝖎𝖙@lemmy.world
            1·
            1 year ago

            This is just a matter of confirmation bias from your side now. You stubbornly refuse to accept factual information very helpfully delivered to you by users who have many better things to do than respond to your inquiries, and you dogmatically refuse to acknowledge that there are advanced and reliable automation tools available for the use case in question. And while you do all that, you belittle the other users in the community by referring to your supposedly superior knowledge and experience, however somehow failing to provide any data or secondary sources to back up your claims.

            Hello Richard, thanks for taking the time to present your criticism. I take some issue with it, but I believe it’s due to a misunderstanding. I’ll explain.

            This is just a matter of confirmation bias from your side now.

            I do think there is some confirmation bias at play here but I’m thinking you may be surprised where it’s coming from.

            You stubbornly refuse to accept factual information very helpfully delivered to you by users who have many better things to do than respond to your inquiries, and you dogmatically refuse to acknowledge that there are advanced and reliable automation tools available for the use case in question.

            This factual information, to the best as I can understand from reading the comments is that these tools exist. I don’t deny that they exist. I didn’t deny they exist.

            users who have many better things to do than respond to your inquiries

            Unfortunately, it would seem you are wrong here, as they have responded to my inquiries. Or, are you talking about the people who haven’t responded to my inquiries? Either way, I agree that there are people who have or haven’t responded to my inquiries that do have better things to do. What is also true is that the people who responded to my inquiries, even if they had better things to do, still responded to my inquires.

            you dogmatically refuse to acknowledge that there are advanced and reliable automation tools available for the use case in question.

            But, I didn’t, is the thing. In the cases where people have provided links to projects, I’ve thanked them. I’ve mentioned my skepticism and I’ve also wished them well on their projects.

            “Advanced and reliable” are marketing terms, and I don’t really care to use them as they have no meaning. Advanced how? In that it uses neural networks? NEAT! reliable how? in that they work 100% of the time? That they don’t generate false-positives or false-negatives? That they don’t degrade the user experience? These are questions worth asking… but… let me be clear: they are questions worth asking for the sake of improving these tools and maintaining the user experience; these are not meant to discourage use of such tools. I believe admins should use all tools available to them, including turning the servers off if they need to - in that toolbelt includes ai based tools and scripts.

            And while you do all that, you belittle the other users in the community by referring to your supposedly superior knowledge and experience, however somehow failing to provide any data or secondary sources to back up your claims.

            Richard, is it, or can I call you dick?

            Either way, Richard, I don’t claim to have superior knowledge. I honestly thought that would come across in my username. Sorry that it did not. I’m a dipshit.

            Like, an actual dipshit. I’m dumber than a lot of people on here. That doesn’t mean that I’m the dumbest, but it doesn’t mean I’m the smartest either. I’m far from being the smartest on this site, and that’s not impostor syndrome. If I sound smart it’s just because I do know a bit on some topics relating to technology and development, and because I have many interests in many topics. What I have at best is a baseline understanding, and I try to remain humble about it. I got pretty emotional in this thread, calling people who disagreed with my very much hard-to-disagree-with stances that “CSAM bad” and “admins should feel free using all tools available to remove and prevent CSAM” names that got me a 3-day ban. A ban which is now ending, allowing me to finally respond to you.

            You see the problem here is that the people I’ve been responding to are people who are misunderstanding some things, which I’m trying to clear up. I believe this is one of the main draw to any message board system, for most people. We like to communicate and share knowledge.

            however somehow failing to provide any data or secondary sources to back up your claims.

            I don’t walk around with sources handy for everything I say, ready to cite them in every single post. I also don’t see many people doing this. I don’t lie, and I’m happy to provide links, sources, whatever - when asked for them. No one’s asked for sources for my claims that “CSAM bad” or that “admins should feel free using all the tools available to them…” or even when I say that AI tools aren’t perfect and lead to false-positives and false-negatives. But, I can give them to you if you want to see them. I don’t talk out my ass. One easy example I’ve been giving people is what’s happening and has been happening for years to this one youtuber, who constantly gets her videos flagged as having a child in them, despite her being 30 and only having herself in her videos. Her experience is valid, as there are some people who will constantly be falsely identified (which again is just stating facts, not suggesting these tools shouldn’t be used - that would suggest that these tools cannot be improved, which they can and are).

            Richard, your comment was a slam-dunk. I’m just not sure you hit the right net. Please let me know if I am misunderstanding anything.

            Yours, Dipshit.