Ben Whitelaw 0:00

Come on then, Mike. So what fun thing did you see on the internet this week then?

Mike Masnick 0:03

Well, this one made the rounds, so I, this might not be something that people haven't seen, but we'll see if you've seen it. Do you know who Mike Tomlin is, Ben?

Ben Whitelaw 0:11

No. Is he, from the '90s?

Mike Masnick 0:14

No. He's, he's an American football, you know, uh, US NFL kind of football head coach for like 20 years. Very famous, very successful as a football coach. And, late last week, he revealed that he spent 12 years of his life while he was the head coach of an NFL football team building the most incredible Minecraft" world. He became obsessed. It was basically the way that he relaxed. He said he started doing it with his kids 'cause his kids were interested in "Minecraft," and apparently he had a long unknown interest in civil planning and how do you lay out cities. And so he started with his kids, and he spent 12 years building up an incredible city, and he released a part one YouTube video where he walks through his city, and it is the most incredible thing you've ever seen. He talks about making sure that there are enough bathrooms for both men and women, and that he's designing... He walks through different rooms, and he talks about lighting and where you place beds and how you have an aquarium and where you should have balconies. It is absolutely incredible, and it is an interesting way in which someone was using the internet to relax and build the most incredible world. I beg you to watch this video.

Ben Whitelaw 1:41

Yeah, I will. I will. and proof that you just never know what people are getting up to online.

Mike Masnick 1:46

Absolutely. Absolutely

Ben Whitelaw 1:48

quite a great use of time when you're not, coaching, NFL teams in your spare time. Fair. Well done. Well done, Mike. Um, my, my thing is, it is about the AI impact on jobs. it's about how we're all kind of impacted in- by this new world we find ourselves in, where anybody can create a CV at the drop of a hat and kind of apply for jobs very easily. It's a game that a guy called Brian Mamaril has created. He's a designer and dev, and he's created this thing called Job Trail, which allows you to, via this kind of game scenario, just see how bad it can be getting a job in 2026 in a world where it, it, you know, anybody can submit, an application at the click of a button, at a point where all of your applications are being screened by AI anyway. he does a great job of like, you'll really like this as a games nerd, giving you options about retraining as an AI expert, taking a break from, you know, the AI news agenda, and just generally kind of like muddling your way through 2026 as you find a job, and I really liked it. I thought it was a great kind of reminder of, of how tough it is out there and, for folks who, who are trying to get a job, it's really tough, and it's a great little game.

Mike Masnick 2:58

Yeah, I'll check it out. As, as someone who loves games that explore difficult topics, that sounds fascinating.

Ben Whitelaw 3:03

Yeah, it's very much you. Hello, and welcome to Ctrl-Alt-Speech, the podcast where we make sense of the major debates shaping online speech, platform power, content moderation, and the future of the internet. It's October the 8th, 2026, and this week we're talking about moderating faster than Sonic the Hedgehog, TikTok and Meta getting serious about child safety, and for our extended Patreon episode, the US administration interfering on an European conference about, you guessed it, foreign interference. Ctrl-Alt-Speech is done in super speed every week by Mike Masnick, founder and editor of Techdirt, and myself, Ben Whitelaw, founder and editor of Everything in Moderation. Mike, it's good to have you back in the chair after spending some, time away at a conference last week.

Mike Masnick 3:52

I, had three speaking engagements in the last week. I crisscrossed the US and was basically on an airplane constantly and then speaking in between each flight. So it was, uh, it was, a

Ben Whitelaw 4:04

Tough. Tough gear, tough gig. Um, well, we, we tried our best to, produce a, a, a worthy episode in our absence, and we had the great Katie

Mike Masnick 4:11

It was great. It was fantastic. I listened on one of my flights. It was wonderful

Ben Whitelaw 4:15

Excellent. And she's, Her book has come out this week. If, listeners, you haven't gone and, listened to the podcast last week, do make sure you check her out, and her great book, "Disrupting Politics" is now available. Mike, we're gonna talk today in the podcast about funding sources, particularly in our Patreon episode. so it makes sense, I think, to talk about how we fund the podcast, and to implore anybody who wants to support the episodes that we produce to do so by becoming either a sponsor, or by becoming a Patreon supporter. Information about becoming a sponsor's on our website. We do, I think, a fairly good job of bringing interesting sponsors and talking to experts in the industry.

Mike Masnick 4:56

We have very interesting conversations with our sponsors that are as relevant content that people enjoy, and we've heard from many people that they enjoy our sponsor segments. And so if you work for a company and you would like to reach the audience of this podcast, you should sponsor us

Ben Whitelaw 5:12

Indeed, Indeed, and if you are not somebody who, works within a company or is not a state-backed entity as we'll, as we'll talk about later, um, you can become a Patreon supporter for either $10 or $20, a month, and that gets you extended episodes of the podcast and earlier episodes so you can listen in almost real time, much like, our kind of furry hedgehog friend that we're gonna talk about in today's episode. So go and find out more via patreon.com/ctrlaltspeech. also rate and review us, guys, wherever you get your podcasts. it really does help us get discovered and, exposed to new audiences and new listeners, so thanks to, for those of you who've done that already. so, have you ever played Sonic the Hedgehog, Mike? Is he-

Mike Masnick 5:53

I I have not. There, there are video games that I've played, and that one missed me somehow. I think maybe I played it once or twice, but, uh,

Ben Whitelaw 6:02

Yeah. I, I think that's fair. I, I had a friend who had one of the early kind of Sonic the Hedgehog games on a Sega Mega Drive, and I used to go around to his house religiously to play. did not play it otherwise.

Mike Masnick 6:12

I think it was one of the games that, just sort of passed me by in terms of like my ability to control it. I think I may have tried it once or twice and I was like, "This is way too hard."

Ben Whitelaw 6:22

Yeah. Yeah. the clue is in the name, isn't it? It's, it's incredibly kind of frenetic, which isn't dissimilar, I think, to, the way that the Discord server was set up this week for the official Sonic the Hedgehog, Discord. we talk a lot on the podcast, Mike, about platform policies and enforcement. We talk a lot about how, companies are not maybe doing their due diligence when launching products or services. the couple of stories we're gonna talk about up front are actually about, you know, users embarking on moderation and, you know, how communities kind of evolve, in, in, fairly quick time. so talk us through kind of these pair of stories, why you think they're interesting and why they make sense together.

Mike Masnick 7:00

Yeah. I mean, I think that, most of the time when we're talking about content moderation or trust and safety, it, it tends to be with sort of, Meta, Instagram, TikTok, YouTube, these sort of big companies. And we forget that there are lots of smaller spaces, and all of them have moderation issues as well. and so this week, there was a, special Discord that was launched for fans of, of Sonic, and it did not go well. And, in the show notes, will have a link to the Kotaku article that details, just how bad it went I think would, would be the, the short version of it. But you have a very engaged group of fans, but many of those fans are, I would say maybe ironic fans, um, that they sort of want to trash talk and want to create havoc and, mess with anyone in any sort of position of authority. So you launch this space that you allow anyone to talk in, and even though there were rules posted, they didn't... you know, the big mistake that lots of people make is launching with no rules whatsoever. There were rules, but, very quickly it appeared that the systems were overwhelmed with people just, making a mess of things.

Ben Whitelaw 8:09

Yeah. Yeah. That's, I think that's fair to say. And the videos in the Kotaku article show just the kind of number of comments, the kind of, the very chaotic nature of it. And, and there were moderators obviously that were in the background getting rid of, rid of stuff that was, I guess, contravened the rules, but still

Mike Masnick 8:27

and banning users and, and trying to take down content. And it was disappearing. It wasn't that they didn't show up with any moderators, but that, those moderators very quickly obviously became overwhelmed, and it was unclear how well prepared they were for you know, this sort of community to suddenly show up and have freedom to do whatever they want and to test the limits, let's say in terms of how they were using the space. But this is the nature of any kind of community, which is that if you are creating an open space where people can communicate, some people are going to test the limits. And, in this case, we're talking about, a large community that is very interested but didn't really have a, an open space like this where they could communicate before. And so it just sort of opened the floodgates, and with it just, an awful lot of nonsense, and is a real clear story of, the challenges of, having any sort of community and, and having it suddenly show up en masse and, and trying to figure out the rules and how do you enforce the rules.

Ben Whitelaw 9:27

Yeah. I, I wanted to ask that, how do you think this plays into the, the impossibility theorem? because you've written a, a lot about how, content moderation is really difficult at scale. We're, we're not at kind of a platform level here. We're not talking about millions and millions of users, but, how does this kind of play into, to that theorem?

Mike Masnick 9:44

Yeah. Well, first I'll start out by correcting you because you, you made the mistake that, everybody makes, which is they say, "Oh, you say it's really difficult." And I don't say it's really difficult, I say it's impossible. And, and that is a, a really important distinction because if it's really difficult, that implies if you just throw more people at it or more resources at it, you can somehow solve it. And there are people, Dave Willner, former guest co-host on, on Ctrl Alt Speech, who believes that with certain technology you can actually get there and make it work at a reasonable level. But the, point of it being impossible to do well is that if you get enough people together, they are going to push the limits, and not all of the decisions are going to be clear. Now, in this case, there were a lot of clear decisions, like clearly people breaking the rules, clearly people just testing to see what would happen if they broke the rules, and that is, part of the content moderation trust and safety experience. I think that, you know, where the impossibility stuff really tends to come in is where you have the sort of gray areas, where it's not clear that it's not, you know, someone clearly breaking the rules or clearly not breaking the rules, but you're just trying to, deal with the, like, well, what do we do here? This is an edge case. This is different. This... There's sort of extra context here. So, some of that comes into this, but I think, with a story like this, when it launches and you're just overwhelmed, most of the focus is on just stopping the madness. And so maybe within that, probably some legitimate users and some legitimate content got pulled down just because, you know, as you're sort of trying to deal with the mass rush, you're, you're probably more in a mood of like take this down rather than take the time to figure out what's really going on. so there are probably some elements of that, but this was, you know... I think this is just sort of like the mad rush part of, you know, of a large community showing up in a place and really... You know, I don't know what it is about, you know, when you have a large group of people who suddenly show up, where some of them immediately wanna test the rules and test the boundaries, but it does seem to be part of human nature.

Ben Whitelaw 11:47

I mean, there is no norms or standards in a, in a new community like this, right? And so

Mike Masnick 11:52

That's true also,

Ben Whitelaw 11:53

o- of the platforms that we talk about who have, you know, robust moderation and who have a kind of civil space as a result, Wikipedia is a, a great example, I think. there are rules and norms and standards that people kind of abide by. in this case, there was rules, but because it's a new space, there, there aren't norms and standards. And, people are coming in as kind of participants in that community rather than maybe having shaped those norms or standards or been involved in that, which is often one of the, the factors I think that I would say help in the process.

Mike Masnick 12:23

I think that's, that's actually a really important point because a lot of like, more long-lasting communities also have sort of, for lack of a better term, elders who are there who help sort of not necessarily enforce the rules, but enforce the norms and standards and sort of tell people like, "We don't do this here." And if it's a brand new space where you don't have that and you don't have anyone who has the respect or sort of a leadership, you know, unofficial volunteer or anything of that nature, and you just have everybody rush in at once. You know, I, I, I sort of think of the examples of what used to happen in the US on Black Friday after Thanksgiving, where people would line up outside doors and suddenly they'd open them and everybody rushes in and people got trampled. I think this was kind of a case like that, you know? This is, you know, they, flung open the doors and they had some rules, but very quickly the, mass of people effectively ignored those rules and started trampling each other.

Ben Whitelaw 13:17

Yeah, I think that's true. I I mean, Sega is a large corporate. Sonic the Hedgehog is a, is a brand. You know, there's a, there's clearly an attempt here to, engage audiences in ways that they feel will help sell products. You know, that's, that's often the goal with these kind of initiatives, right? And Discord has been chosen as the place for that. in that regard, like, there are many, many examples in history of, companies and organizations trying to do this, right? Create spaces where players in particular of, games can come together and, and do this. and, you know, big companies like PlayStation and Xbox have had forums for many, many years that most of which are being shut down now because of, difficulties managing those and, and them getting out of hand. But to what extent do you think, like, this is a thing that's kind of r- ending or, or running out of steam? This idea of, like, branded communities where, an, like Sega decides the rules and, and asks people to come into it, and then kind of tr-tries to, I guess, sell into that audience?

Mike Masnick 14:11

Yeah. I mean, I think it sort of continues to go in and out of style, right? I mean, lots of companies wanna keep sort of the brand recognition up and wanna keep people engaged and see, future opportunities to make money. That's what these companies are, are doing. And so building communities is kind of a natural thing. It happens all the time. It's just a question of how do you do it and how do you do it well, and recognizing that you can't just set up a space and say, "Have at it." You need to actually have some element of enforcement. You know, in this case, they did, and I think it's important to say we're not saying that they just flung open the doors with no, no plans. They had rules, and they had some, moderators who were there to enforce. It was just that they became very quickly o- overwhelmed. And you sort of have to recognize that if you're building a community, that you have some responsibility there, and some responsibility to figure out how do you manage that community, how do you govern that community. And that is sort of like the theme of this podcast, I think, you know, from, from the very beginning, is that people will gather online, and whoever is setting up those spaces has some responsibility in terms of how it's set up. And this was just, I thought, a really interesting example outside of, you know, it is using Discord certainly, which we talk about, but This isn't a Discord corporate issue of, of how to moderate it, but how to moderate, a popular but smaller space within a Discord community. and it, it just sort of shows how at every level of the stack, there's some element of content moderation or trust and safety that people need to be thinking about. Which I, I think actually leads us to the, the second story that we're, we wanted to connect to this, which was there was a fascinating experiment earlier this year where, three other tech journalists and podcast hosts. This is,

Ben Whitelaw 15:51

Wait, we're not the only ones?

Mike Masnick 15:53

are not the only ones, but three other journalists and podcast hosts, set up their own Mastodon instance. And so this was Casey Newton and Kevin Roose, who at the time had "The Hard Fork" podcast, at "The New York Times," which they just recently left, and they've set up a, a new thing called "Machine Gods," which will be launching in a few weeks, I think. along with PJ Vogt, who has the wonderful "Search Engine" podcast, and has been a longtime podcaster, uh, had done "Reply All" and a podcast called "TLDR" before that. and so he has the "Search Engine" podcast, and, and the three of them got together and they were talking about social media and how much they hated social media, and it was no fun anymore. And I forget who decided that why don't we try to set up our own Mastodon instance, which they eventually named the Forkiverse. And so that launched at the beginning of this year, and it got a lot of attention, and people were really excited about it. and people were saying it was such a nice community and so much fun, and I, the cynical veteran of seeing online communities form and devolve over the years sort of felt like, look, almost every community when you first start it out is nice and fun, then it becomes a real mess.

Ben Whitelaw 17:08

E- even the TechDirt comments?

Mike Masnick 17:10

Uh, you could say that. That's one example of it. Though, you know, I think we do a pretty good job of, uh, enforcing our, our setup in the Techdirt comments. Uh, we've put in some incentive structures but we heavily monitor what is happening in that space, even if we're relatively free in allowing, even bad actors to engage in some, some bad activity with comments. but it is, kind of a challenge, that it is something that you have to sort of constantly watch over and weed through. And what happened, what we found out this week is that one, Russian trolls had invaded the Forkiverse and sort of taken it over, that the three people who put it together, who are all very busy with their day jobs, had more or less, abandoned the community, were not keeping close control over it, were not monitoring it, and weren't aware necessarily that Russian trolls had started to take over the space. And they've now announced that they're going to be shutting down the instance altogether by the end of the year, and they're asking the remaining people who have been using it to start looking at transferring to other, Fediverse instances, which, you know, is a very nice thing that you can do with ActivityPub and Mastodon and the Fediverse, which is that if your instance is no good or if it's shutting down, you can move to a different one. And I think that's, you know, part of the, the wonders of more decentralized systems like this. But I do think that there is an interesting story there in that it is really fun to set up a space and say, "I have a community," and to bring in people and have discussions, and that part is exciting. And then the real work begins, and you need to prepare for that, and you need to have, again, rules and enforcement and a process in place to do that. And, you know, this is an example of what happens to many communities. I'm not saying that, these guys made any mistakes at all, but, like, this happens to so many communities. People start a book club on Facebook or, you know, whatever it might be. People begin a community and it is fun and exciting in the first, days, weeks, months. And then at some point it begins to become more of a chore, and part of that chore is setting up the rules and then enforcing the rules and dealing with, say, Russian trolls or whoever else decides to show up and cause trouble in your space. And it is this reminder that trust and safety and content moderation are not just things that Meta and YouTube and TikTok need to take care of. They also happen in all sorts of smaller communities as well.

Ben Whitelaw 19:40

And, and increasingly I think, you know, everyone will become a moderator. I mean, uh, uh, not even increasingly. I think this idea should be sharing more widely, is that everyone is moderating their own experience right now. You know, this, this is completely lost on a lot of people, is that there are ways and means, even on the big platforms that don't offer a lot of controls of you blocking people, of, exiting groups, of you even shaping the, algorithm that you use by liking posts, by hovering over videos. There are, They are all signals that the platforms use to show you content and it's not just about content that is removed, it's about content that is reduced, increasingly. And so I think this story or these stories show just that really that everyone is a moderator, and that we need to kind of get this idea to people that everyone is moderating their experience online. You can do that in lots of different ways, and in this case, where these excellent tech journalists have, gone to create a new space. They've, they've not done a great job, but it's, it's

Mike Masnick 20:44

mean, they they did a great job at first, right? they started a great space that people really liked. that is the thing that is important, is that it's great to create a space, but to recognize that, maintaining a space takes work. And you're right, like, everyone is sort of a gardener of their own internet experience, and some of that involves pulling out the weeds, and that can be blocking people or deciding who you follow or deciding, you know, exactly as you said, which things to click on. That is, part of the process of gardening your own internet experience. But, like, I think it is really important for people to realize that, you have some responsibility for your own experience, but you also have responsibility for the experiences of other people. when you create spaces, you have some responsibility there. And, I think that's an important lesson that a lot of people don't ever think about, or they expect that somebody else is supposed to take care of that. And some of that is that, you know, we've trained people for so long to think that, like, Meta or TikTok or YouTube are in charge of policing these things rather than recognizing that we have some agency and some responsibility ourselves as well.

Ben Whitelaw 21:47

Yeah, agreed. And, and, you know, respect to people who, do that in their own internet gardens, right? The, subreddit moderators, the, stream moderators, the p- everyone who's doing this kind of work in their own little ways on local news forums or, as you say, YouTube channels. Like, these are all important contributions to the web, and they're, and they're difficult to get right. And, you know, even in my own experience trying to do that on a news website with a few thousand commenters every day was incredibly, difficult to, you know, thread that needle. And, the fact that these journalists have, have tried, initially done well, and then kind of fallen, or fallen foul of it, seen how hard it is, I think is, proof of, how tricky it can be.

Mike Masnick 22:27

Yeah. this is not a unique story. I mean, we're, we're not like, we're using them as an example. We're not calling them out as if they made a mistake, because I think lots of people discover this. Lots of people create spaces and then, end up ignoring it because, you know, it's fun when you're first creating it, and then it, there's work involved. And, you know, sometimes that work is empowering as well, and sometimes it's, you know, it has its own benefits and its own rewards. but, you know, I think it would be good if more people realize as they're creating online spaces that there are some responsibilities that go along with it.

Ben Whitelaw 22:59

Yeah, indeed. well, there's, I think there's a, a moral there, Mike, which is which I hope our listeners, will understand. we're gonna talk about a couple of other stories that caught our eye this week, which are kind of loosely bound in big platforms trying to do child safety better, I would say. we're gonna start by talking about a really good piece from UK magazine Prospect, which has, uh, just done a kind of investigation into child sexual abuse material on TikTok, and essentially over a period of kind of months and a lot of searching, found 176 accounts with best part of two million followers who are acting as kind of gatekeepers to child sexual abuse material that is being traded and sold via messaging apps. some of what we're gonna discuss now is graphic and fairly distressing, so just to kind of give listeners a bit of a warning there. we won't go into lots of detail, but I think it's important to, explain the story and, and the piece. I mean, this, is not new in lots of ways, Mike. We know that platforms, in some form struggle with, being able to detect child sexual abuse material, and there are lots of, there's lots of work being done by a whole bunch of technology companies and civil society orgs to try and prevent that. so in many ways it makes sense that there are some accounts on a platform as large as TikTok that have crept through. You know, it has all the hallmarks of, other examples like this. users using hashtags to try and spread the word about the fact that they trade this material. they use kind of algo speak to avoid detection, so they use, words that can't be, detected by some of the filters or technologies that these platforms have. And so in many ways, that's not what was kind of interesting about it to me. What I found really interesting was the fact that a subset of the accounts that they found, about 20% of them, used an image of a ginger man, a red-haired man,

Mike Masnick 24:49

This was the part that stood out to me too, yeah

Ben Whitelaw 24:51

Right. And, And, again, this just shows how harms evolve and how it becomes increasingly difficult to detect harms, because this red-haired man, the journalist found out, is a man called Matthew Estes. And the various kind of images of this man are AI generated, but they were able, able to trace it back to a photo of him. Matthew Estes is a 27-year-old man who is currently in, prison in the US for 60 years for the production of child sexual abuse material. And I'm not gonna talk about the actual crime he committed, but he himself had a very troubled upbringing, was involved in the production of child sexual abuse material, and is now being used essentially as a signal to other accounts that they sell and trade this material. and it just kind of shocked me, Mike, the extent to which, you know, how difficult it is for a platform to know that that image of a, of a red-haired man is a sign of child sexual abuse material. and I'm sure, once a platform does know, they're able to do something about it. But, that was just the killer detail for me and, quite shocking, I think

Mike Masnick 25:56

Yeah, I had the exact same reaction. There's a lot-- This is a really well-reported, detailed piece. I highly recommend it to people who are interested in how difficult it is to kind of track down this material, and how widespread it is. but the one thing that really stood out was that detail of, you know, we knew that people try and hi- obviously try and hide the fact that they're, selling or offering child sexual abuse material. but I'd never heard of this method before of basically creating, using AI to create an image that is not CSAM, but is just of this individual, which is used as a signpost to say that I am engaged in the, process of, selling CSAM, and then pointing you off-site to another platform, right? So there's layer upon layer of showing the difficulty for any platform to try and figure these things out to, one, have to know that, not the same image, but AI-generated images of this particular individual is being used as a signpost that people in that community recognize and know that they can then follow that path to another site that TikTok doesn't have control over and recognize that then they're, they're engaging in the, the sharing or selling of CSAM.

Ben Whitelaw 27:10

And that's often less well moderated, right? And, and harder to, to moderate. So Telegram is, the example in this piece, but it could be any kind of messaging app frankly. And a lot of those messaging apps are encrypted to some degree or in Telegram's case have, varying degrees of content moderation processes in place. So it, it's not surprising I think that TikTok said that a lot of the activity that was flagged by the journalists actually didn't violate its policies because you can't say that a, picture of a man is violative in and of itself

Mike Masnick 27:40

Right. This is part of the problem, which is like, how do you handle... You know, this has always been an issue with, trust and safety stuff, which is, like, how do you handle off-platform behavior if all of the on-platform behavior is perfectly reasonable? And, what's happened here, and obviously I'm, I'm assuming that this is not the only picture that is out there that is, that is a signpost. I'm sure they have their systems of, you know, letting other people know what, signposts work for this kind of thing. But on TikTok, the content appears perfectly normal and natural, and there's nothing, inherently suspicious about it, which demonstrates the difficulty of a big platform like TikTok and trying to have to track this stuff down. you know, if somebody's posting totally innocuous content and then people report it, they're gonna say, "This doesn't violate our rules." And to have to know the multiple layers deep, and some of it totally hidden, that it's actually leading to CSAM, it just, you know, raises the difficulty level, raises the impossibility level to just give you a suggestion of what a ridiculously impossible task it is to track all of this down and stop it. and I appreciate the reporter who went through all of this work to, like, piece these things together. And, you know, it would be great if trust and safety teams had the resources to do this over and over and over again. But, you know that the people who are engaged in, in trafficking this stuff, have a long list of ways to get around, being spotted. and it just sort of really tells the story of how difficult this is. I will note, related to this story, uh, soon after this story came out, there was a separate story about Meta in particular, that they've released a new AI tool which they claim is very good at tracking down these kinds of things, in terms of figuring out when people are pushing people to off-platform, situations for the exchange of CSAM material. They claim that they're much better at now tracking that down with a new AI tool. I hope that's true. we don't have any way of testing the actual, ability of those tools. But it is a sign that at least, some of the companies are obviously aware that this is happening and are looking at ways to use the technology to prevent it.

Ben Whitelaw 29:50

Yeah, and I will just say that that, announcement is following pressure that came about as a result of reporting again, about this issue in India, which found that there was ads that were being, paid for by individuals that were then pushing, other users to go to Telegram chats and, either trade or buy child sexual abuse material. So I think what we're seeing here is a, renewed focus, Mike, on CSAM as a, kind of harm factor as a result of reporting like this Prospect piece, which I think is, is only a good thing. You know, there are questions about why, for example, at Meta, it's taken them a bunch of time to create this tool, and why it wasn't done previously. But again, I think this is, the nature of the push and pull of media coverage of these harms and platform priorities, frankly. So yeah, I think good to kind of knit those together. You also wanted to talk about in a, in another eye-catching story that were gonna include in the show notes, another TikTok story related to, again, a, something we've talked about on the podcast previously. so trying to do our due diligence here by following up on stuff we talked about, Mike. Um, talk us through this one.

Mike Masnick 30:54

Yeah. So, we've seen a lot over the past few years about complaints about the different tests that companies run in terms of, safety and, and offering up A/B testing and different things. And sometimes those are portrayed in the media and by politicians and by law enforcement as, clearly negative. Like, "Oh my gosh, how dare they, test these safety features or turn off these safety features for certain people?" and I've argued against that, and I've said, This is how you find out what works." You know, the company should be doing tests and should be seeing what works. and that includes TikTok, which has been accused of, sort of pretending to offer safety features when they knew they were turned off or whatever. But this latest example, which is coming out of the lawsuit that New York filed against TikTok, seems a little bit more extreme, and I think probably crosses the line from, like, reasonable A/B testing or control testing into something that I think looks really bad. And so the specifics that came out, and again, this is one-sided, this is from New York's side of the lawsuit, and TikTok says they've denied it, but we haven't seen, like, the full filing in terms of what it actually means, is that TikTok announced specific safety features to sort of reset your algorithm and encourage teenagers to use it. except that they ran a test where if a teenager reset the algorithm, it didn't work, and they just lied. So that is not a good test. Like, I can understand a test where some people have the option to use this tool and some people don't, and then you have a control and you see how that works. But telling people that you're giving them a safety tool and it will reset your algorithm, and having them click on it and do that and think it's happening, and then it not happening because you're running a sneaky test where you have told them something is happening that doesn't, that to me crosses a line.

Ben Whitelaw 32:47

it's gaslighting, isn't it? you're gaslighting your own users es- essentially. and not only that, but putting them at risk while you do so

Mike Masnick 32:53

Exactly. Yeah. it's one thing, and I know, like, there are other cases where, because a company in their sort of marketing material says, "We take safety seriously," and then somebody gets hurt, and so, they get sued and they say, "Well, you lied when you said we take safety..." Like, those I think are a little bit nonsense, right? Because you can't have a perfectly safe system. There's always going to be certain challenges, and I think that is unfair. But to deliberately lay out a tool and say, "We are giving you this tool for your safety," having people click on it, expecting that it will do something to improve their safety, and then having it do nothing on purpose, that strikes me as that, that is ethically dubious and just sort of, directly dangerous to the people who thought they were getting a safer experience because of it.

Ben Whitelaw 33:35

Yeah. And at a time when product safety, is such a focus for the industry, right? the cases brought against Meta in the US have focused a lot on product liability and safety design, and the ability to kind of build in features and functionality that proactively preempt harm. is a real kind of, I think, case to say that historically, platforms haven't done that as well as they, could or should have done.

Mike Masnick 33:57

Yeah. and look, it is challenging, and again, like, I support the companies doing tests and figuring out what actually works and experimenting, and I worry about a lot of the lawsuits and a lot of the media coverage that implies that any sort of testing itself is sort of acknowledgment of harm. but this is not that. This just feels bad.

Ben Whitelaw 34:16

Yeah. this is, um, an error of judgment, I think

Mike Masnick 34:20

Yeah.

Ben Whitelaw 34:20

is,

Mike Masnick 34:21

someone messed

Ben Whitelaw 34:21

put it. Definitely. Definitely. and Before we move on to kind of our Patreon segment, we wanted to nod to a story that one of our Patreon supporters and listeners, has also flagged to us. this is a kind of, like, fairly wonky story. It's, it's technical, so I'm gonna let you get into this one. Um, I know that's your bag. but it's interesting for reasons, that we spoke to on the podcast a few weeks back

Mike Masnick 34:42

Yeah, and we won't go too into the weeds on this, but this is about a company called Quad9 that does DNS resolving. We talked about DNS resolvers and how, when there's sort of attempts to censor speech on the internet, people, companies, governments, including the US government in the case we talked about a few weeks ago, look for sort of the choke points, and that can be sort of deep within the network stack, and that can affect the speech of many, many people. In this case, we're talking about, Quad9 that does DNS resolving, and they faced issues with the oldest... Well, maybe not the oldest, but one of the oldest content moderation questions that comes up having to do with copyright infringement, and the copyright holders who want to block sites entirely that are engaged in sort of, some sort of infringement, and they've been targeting deeper and deeper into the network stack. And in this case, in France, they're trying to get the DNS resolver to block access to certain sites, and Quad9 has refused to do that and basically saying that's overkill, to block access to sites would sort of be a global censorship regime. and they're unwilling to do it, and now they're talking about moving out of France entirely. so it touches on a bunch of really important points. It feels sort of wonky and technical because it's deep within the network stack, but I think it's a really important moment in how, internet speech works and how we have these intermediaries that we rely on deep within the stack. And the deeper you go within the stack, generally the impact area of if they actually do anything is much larger because it impacts, you know, it's not narrowly targeted in any sense. So I think it is an important story, and we'll throw it in the show notes.

Ben Whitelaw 36:15

Definitely. And, and for listeners who want us to talk through any particular stories that they've seen in their, their feeds or come across their desk, remember, if you become a, an insider supporter of Ctrl-Alt-Speech on Patreon, then we will, we will happily, assess those stories, talk about them on the podcast, and share them with our, our audience. So, uh, remember you can do that. Mike, we're gonna transition now to our Patreon extended episode. talking of choke points, talking of, uh, y- know, kind of dilemmas as to, what to say. I don't know where conference events sit in the stack, uh, but, but I feel like it's kind of not particularly deep or technical. this is a story that we've seen across our feeds over the last three or four days. it's about a conference that's happening in, in Lithuania, which has a whole tranche of disinformation researchers, who are there to talk about their work, to share their research, but who've actually fallen foul of some pressure from the US administration. I think it's fascinating for loads of reasons, talk us through why, why you think it's worth covering.

Mike Masnick 37:14

Yeah. I mean, obviously we talk a lot about disinformation and we talk a lot about the research and efforts around how do you respond to disinformation, how do you handle counter-speech. you know, a lot of the people at that conference are people that we know or that we've spoken about or that we've read their research or all of those sorts of things. and, you know, there were some odd goings on that basically came out earlier this week. and then, as more details came to light, it turned out that the US government was deeply involved in effectively trying to interfere in this conference and block, or at least condemn a certain panel that was happening that, that asked whether or not the US would interfere in, foreign information environments. And so I think it touches on a whole bunch of the topics that we talk about, including disinformation and government attacks on speech, and so I think it's worth digging in as to what really happened and why it's such a big concern.

Ben Whitelaw 38:10

Yeah, we'll do that now. Thanks, mate

Announcer 38:14

That's the end of this week's free edition of Ctrl Alt Speech. To hear the extended episode with additional coverage, become a Patreon supporter at patreon.com/ctrlaltspeech. That's C-T-R-L Alt Speech. If your organization is interested in sponsoring the podcast, contact us via our website at ctrlaltspeech.com. Thanks for listening