Ben Whitelaw 0:00

So Mike, tell me, tell me, tell me, what fun have you had on the internet? What fun have you come across? What made you smile?

Mike Masnick 0:07

So this one is fantastic. this is something called Chipotle Max. instead of the E at the end of Chipotle, as the, restaurant chain is called. Do you have Chipotle in the UK? Is that

Ben Whitelaw 0:19

Um, I don't think we do.

Mike Masnick 0:23

Okay.

Ben Whitelaw 0:23

Certainly not in London where I live.

Mike Masnick 0:25

Okay. It's a burrito place, and it's, you know, it's good. You should have it next time you're in the US. Uh, I mean, they had some trouble a few years ago where people were getting sick eating them, but I think they fixed that. but anyways, this is Chipotle, but instead of an E at the end of Chipotle, it's AI. Chipotle Max. And what happened was somebody discovered that Chipotle, the restaurant, has a support chat that is powered by AI on their website, but they left it exposed, let's say, such that people could get the Chipotle customer service bot to respond to other things and give you free AI

Ben Whitelaw 1:03

Oh, okay

Mike Masnick 1:04

wanted it. And so they built a tool that taps into that and allows you to effectively get free AI, uh, power, free AI tokens, thanks to the burrito shop.

Ben Whitelaw 1:17

Bu- bu- buy a burrito, get some AI access for free. I think that's a fair deal. It seems to be the way the world's going right

Mike Masnick 1:25

Yeah. And they do warn you that, you know, this could be fixed and Chipotle could close it off, but they're in the process of trying to add AI support from Home Depot, Nordstrom, Sephora, Ikea, and many others. So you may, may get free AI from all of these customer support chatbots that are not properly secured.

Ben Whitelaw 1:45

So have people been building stuff with...

Mike Masnick 1:48

Yeah. Yeah, it's, it's just out there.

Ben Whitelaw 1:51

Y- have you been vibe coding with the Chipotle

Mike Masnick 1:54

I, I have not. There, there, there, there are limits to the things I'm willing to experiment with but it's, uh, I thought this was absolutely hilarious

Ben Whitelaw 2:04

Yeah, great. And, uh, speaks to some of the stories we'll talk about today.

Mike Masnick 2:08

Yeah. What about you? What fun thing did you find?

Ben Whitelaw 2:10

well, after you alerted me to 82 and Zero, that very fun basketball game, and I've got more into basketball over the last few weeks, as we might touch on. inevitably there is now a World Cup version of that game,

Mike Masnick 2:24

Of course

Ben Whitelaw 2:25

it's called Seven and Zero. we'll leave the URL in the show notes if you want to kind of, spend a lot of time over the next few weeks while the World Cup is on trying to beat this game. But essentially, yeah, you, you have to pick a squad, from all of the years of the World Cup and, uh, yeah, get seven games in a row. And I haven't managed to do it yet.

Mike Masnick 2:43

you sent this to me yesterday and I was playing around with it, and I am not good. I also know very little about World Cup football. But, yeah, I'm not good at it.

Ben Whitelaw 2:52

No. Well, by the time this World Cup, which is in your, your fair nation, um, at least some of the games ends, maybe that will be different, and, uh,

Mike Masnick 3:00

I'll pay attention

Ben Whitelaw 3:01

you'll be breezing through this, seven and oh game before long, so, yeah. Happy World Cup Day to those who celebrate. Hello, and welcome to "Control 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 June the 11th, 2026, and this week we're talking about Apple's increasingly large role in the safety stack, the reaction to the oversight board's big announcement last week, and for our Patreon subscribers, why Anthropic's powerful new model hasn't had the easiest of starts. Control Alt Speech" is the internet version of Monday Night Football with with Mike Masnick, founder and editor of Techdirt, and me, Ben

Mike Masnick 3:50

These are getting better and better.

Ben Whitelaw 3:52

founder and editor of Everything in Moderation. you know the thing about that reference, Mike, is that it speaks to both your version of football and my version of football

Mike Masnick 3:59

Okay. You have a Monday Night Football in, in the UK as well?

Ben Whitelaw 4:03

there's a soccer Monday Night Football, which is, as you might expect, pundits, talking around the Monday game of the Premier League. And I also found out when I was preparing for today's podcast that you also have that, but for, for football.

Mike Masnick 4:17

Yeah, for NFL football. yeah. No, it is, it is one special game each week. Most games are on Sunday, and they have one, unique game that is on Monday night, and it is a, a tradition in the US.

Ben Whitelaw 4:30

I think the Premier League, the, UK Premier League stole that from the NFL. We didn't always have that. Now we do, um, which means we can use that reference. Um, we are up against it in loads of ways today, Mike. we have so many stories that we're trying to kind of cram in today's episode. we're also trying to finish before 8:00 PM UK time so you know, at least I can watch some of Mexico versus South Africa,

Mike Masnick 4:54

There we go.

Ben Whitelaw 4:55

which is the first game of, this year's World Cup. you mentioned you're not a fan, but are you, are you gonna be tuning in? What's your, what's your plan for the next few weeks?

Mike Masnick 5:03

I had not directly planned to tune in, but I am sure I will. I'm-- Right now I've been pretty focused on basketball, uh, because that's where my, my... The NBA finals have taken up my sports focus right now. But I'm sure I'll start to, start to tune in to the World Cup games as well

Ben Whitelaw 5:20

Yeah. As soon as the Knicks win the next game, right?

Mike Masnick 5:23

Yeah, they just need to win one more,

Ben Whitelaw 5:25

One more game. Um, we, we were talking before we started recording about the big result last night.

Mike Masnick 5:31

Yeah, it was fantastic. The best basketball game I think I've ever watched in my life.

Ben Whitelaw 5:36

High praise. High praise. I mean, if, if the World Cup can, get anywhere close to that, it will have done well. and I enjoyed watching the highlights now that I am, an honorary Knicks fan. it's already happened. we have loads to get through, so I'm not gonna dwell too much on, sport, but we wanna say a big thanks to our burgeoning group of Patreon subscribers, the OGs, the founding tier members who, are kind of crystallized, memorialized. you can no longer become a founding member, but that status is locked in for life for those people who signed up on that tier. We've got our supporters who pay $10 a month to get extended post-match analysis, as, as I'm calling it now. And, uh, and also our insiders, is the new version of the founding tier, who will get longer episodes, but also get a direct line to Mike and I to suggest articles that we missed, where we went wrong, you know, VAR calls and decisions that perhaps we, we got wrong. So go to patreon.com/ctrlaltspeech for that. we are gonna start actually, Mike, with a story that was suggested by one of our Patreon supporters. and, we're gonna expand on, on that. Our first big story really looks at child safety announcements made by Apple this week. But it kind of begins somewhere in the middle of last week, with the ongoing dance between politicians who are clearly keen to win votes and big tech companies who many voters have taken a keen dislike to. So one of our Patreon supporters, Joe, highlighted a story last week that was previewing a speech by my fair Prime Minister, Keir, Starmer, in London Tech Week this week. And, Keir, this article kind of highlighted, was going to insist that Google and Apple would do more to keep kids safe. And the focus of that particular story was the fact that Google and Apple execs could face prison time if they didn't do so. And I know you love stories like this. I know you've got things to say. When, when this story kind of made it into our Signal chat, what did you make of it?

Mike Masnick 7:41

Uh, I mean, it's just, it's the same old thing. and it's very much-- And, and that original article came out before he'd given the speech, and then he, he gave the speech, and it was basically, it was the exact encapsulation of the nerd harder problem, which was, these companies better wave their magic wand and not allow anything bad to happen on their platform or they're going to jail. And like, it's just so unrealistic, and it creates such twisted incentives where not only are you asking for the impossible, you're then threatening, literal prison time for not doing the impossible. That doesn't create good incentives for actually building better systems that are actually safer for children. And so I find this, very performative and showy and, not particularly helpful or realistic

Ben Whitelaw 8:30

Yeah. yeah, there's are reasons why politicians use the prison time, threat. we won't go into the kind of depths of that, but, um, yeah, it, it doesn't create the right incentives. I think as you say, the speech actually took place, this week on Monday, and the thing that was reported very widely was a three-month deadline for Google and Apple to prevent, nudity, on those platforms. So, so they basically Starmer said, "You've got three months to ban nude images on children's devices," as part of an effort to keep kids safe. And for those of you not very close to kind of UK politics right now, Keir Starmer is under a lot of pressure. there is a, guy called Andy Burnham who is standing in an election, a by-election next week, who people are thinking might be the next prime minister. So Starmer is in kind of survival mode, and one of his kind of survival efforts is to bolster his child safety, internet safety credentials. And so, that was a really interesting moment in itself. That was gonna be probably our big story, and we were gonna kind of unpack some of that. Then at the same time, Mike, on the same Monday, Apple did their big developer conference in Cupertino. And guess what? They checkmated Keir Starmer. Um, or at least I think so. Um, and I'm keen to, hear what you think. Just to give you an overview, for those who aren't staying close to the kind of the tech news scene, this is the WWDC, the conference that, Apple hosts every year, normally does a kind of product showcase explaining all the great features and functionality that it has launched for, its developer community, but also for users of Apple products. It was a big one because it's Tim Cook's last one, but it also was a big moment for Apple taking kind of center stage in terms of its safety efforts. it is very much worth watching at least the 10-minute clip about trust and safety in its entirety. we're actually gonna play a bit for you now just to give you a sense of what it's like. so take a listen to this and see what you think.

CLIP: Craig Federighi 10:31

Our long-standing commitment to safety is especially important when it comes to kids and teens who rely on our devices to stay in touch, to explore creativity and enhance learning, and to build their independence as they grow. And as parents ourselves, we're committed to building a safe and trusted platform for kids. So this year, we're taking a big step, expanding our child safety features with powerful and intuitive tools

Ben Whitelaw 11:03

So that is a clip from the Apple presentation. You hear some of the trust and safety executives talking about what they're planning to do. there's a whole bunch of features so I'll just kind of go over them. You know, pretty significant changes to how they do safety. The, the biggest one I thought was the Ask to Browse feature, which builds on what they do in the App Store already, which is, allowing children to request access to specific websites, via a parent. So to kind of essentially gatekeep, particular websites that parents might not want children to access, which is maybe quite burdensome for parents, but it's probably gonna give them a lot of comfort and, feeling of safety. they've also expanded the, what Kiera Starmar has essentially asked for, which is they already have this, ability to prevent nudity on messages and FaceTime calls, and they've expanded that to also contain images that reflect gore or violent content. So they're gonna be blurred now as well on the Messages and FaceTime apps. And then they did a big also announcement about screen time. screen time is a big issue for, parents. they've made those controls much more granular, so you can now specify not just kind of screen time overall, but entertainment apps, games apps, and social apps now can have particular, time allowances. And you can schedule different time allowances for different days should you wish as well. So if you wanna allow your kid to watch a film on the weekend, but you don't want them to use certain apps on, on a Tuesday or a Wednesday after school, you'll be able to do that within Apple's safety features. So it's a big moment, I think, in lots of ways, not least because, politicians have, gone after Apple and Google in trying to, get them to do more. But also because, you know, Apple are gonna take a, an even bigger role in the, the trust and safety stack, Mike, which is, we'll talk a bit about. What was your assessment of, of this, big announcement from them?

Mike Masnick 12:53

Yeah, I mean, there are a bunch of different angles to this. so one which you sort of alluded to, but there, there was an interesting article on, in 9to5Mac that actually called this out where, the normal WWDC presentation sort of walks through the different operating systems and the different, systems that Apple has and sort of talks about new features for each one, whereas this one had like this entire section on trust and safety, which is new and stands out and is unique and, and certainly interesting. and so that is on its own worth noting and, calling out and the, sort of central focus of trust and safety, which obviously we think is important. So I, I sort of lump the announcements here into sort of like the good, the bad, and the ugly uh, categories because there are different elements to this. You know, one, the good, the features here that they've talked about and safety features that I think are very good, and I think Apple in particular has, has always done a pretty good job of thinking through safety features and actually thinking through them, for the most part, in a very thoughtful way that are privacy protective and that really do help and are useful in, protecting children and you know, being, being actually useful. And so I appreciate that, and it's, nice to see sort of the latest generation of those. the bad part where I get a little bit more nervous about it is the idea that it might start to lock in Apple as sort of essential infrastructure for child safety. This is my concern where you have this one large company that becomes such a central piece of it. If, if we start to look at the operating system providers as having to be the sort of controllers of all of the safety features, that puts a lot of power in their hands. And, while we can say that, while we can hope that Apple is a sort of benevolent dictator, those sorts of things tend to get corrupted over time, and it becomes a riskier and riskier situation when they are sort of locked in and they have too much power. And so this is, you know, what some of you know, the essay that I published last week that we spoke about, becomes a a potential issue. And then the, the sort of ugly end of it, I think, is there is an element of this, and as with everything there are, there are trade-offs that I feel sets this standard of surveillance that we're sort of training our kids that surveillance is the norm and that they should expect to be surveilled at all times. And I think that itself is somewhat unhealthy. And I think, obviously there are different levels to this and different, ways to think about it. And obviously for younger kids you want, parents looking over their shoulder or trusted adults who are teaching them how to use the tools appropriately and in an age-appropriate way and all of these kinds of things. But I worry about a world in which we sort of bake in this idea that you are going to be surveilled until you're 18 and then we suddenly let you loose in the world without, the surveillance. I think there are longer term consequences to that that are really problematic, and that, if we sort of accept this position that this is the way it is, everything that you do and every conversation that you have is, being tracked and monitored you know, every app that you access, all of that kind of stuff, over time that is not a particularly healthy situation either.

Ben Whitelaw 16:09

Mm-hmm. w- let's kind of start at, the ugly end of that spectrum. What, what, what are you kind of most concerned about? Or where do, where do you see that kind of surveillance going? I mean, these changes aside, as we all know, internet users are surveilled in lots of different ways, for purposes of being served advertising as the kind of probably most egregious. what is it that this particular type of, I guess, kind of profiling might be useful?

Mike Masnick 16:35

I mean, there's a few different elements to it, but, but part of it is just sort of baking it in as if it's necessary, right? Like, the other kinds of surveillance aspects that you're talking about There are things that users can do to avoid, right? There are steps that you can take to avoid those or to limit those or limit the impact of it, and it's something that there are other tools out there. This is more baking it in as, like, a core feature that has to be there in order to use these tools. If you're a certain age, it is expected that you are going to be, watched in all sorts of ways. and, I just sort of worry a little bit about the kind of message that that sends and the kind of expectations that that leads to, especially for younger people that, Actually, it is important for younger people to, have their own thoughts and, and have freedom. I mean, this goes to a, different extent, and again, like I talked about this a little bit in, in that piece that I had written last week, the idea that, Danny O'Brien had put together of, cognitive liberty, that when everything that you're doing is controlled by centralized parties who can see what you're doing and track what you're doing, it actually limits your ability to think freely, and that itself is a problem. And I'm not saying like, "Oh, you know, nobody should ever, you know... You can't, can't have any other restrictions or guardrails or all that kind of stuff." Again, these are all about trade-offs. But the idea that we're baking it in entirely that everything you do is going to be monitored and tracked and watched. And again, there's, like, an implicit as- assumption when we're talking about parental controls in particular. There's an implicit assumption that, parents know what's best for their kids, and that is not always true, and there are obviously kids who have poor relationships with their parents. there are estranged, kids who are estranged from their par- There's all sorts of things where these things could go wrong as well. And yet, if we're giving those adults the power to, track, monitor, and limit every little aspect of what that child does, I can see where that can lead to harms as well. And again, there are obviously situations where that's not true and that it is helpful to have that, and I do think, like, tools for parents and kids should be this sort of, like, ongoing negotiation of what is appropriate and what is not. But I worry when you have one company that starts to take over, and it becomes this sort of thing that the government is expecting them to do, and you have fewer options in terms of how you, how you lay out those, tools and how you use them.

Ben Whitelaw 18:55

Yeah. I mean, that's a, it's a really good point. I mean, I'm not gonna play defense for Apple here, but one of the points I kinda noted was, was how people have been calling for Apple to play this role increasingly, and I think it's worth laying that out for listeners. So, I mean, there's, to begin with, Starmer's comments about Apple and Google taking a, an outsized role in the kinda safety stack and doing more to, protect children is, in my mind, as a result of a lot of parenting campaigning groups pushing for a stronger kind of online safety regime and, that I think has led, whether they knew it or not when they started pushing politicians to make these changes, into, a situation where politicians like Starmer are pushing Google and Meta and Apple to do these things. So I think there has been a push from, I think the desperation of some parents and I'm hoping to speak to some of those i- the future on the podcast, kinda led to this. That's the first thing. The other thing is that there have been trust and safety experts and people working in platforms who've also, over the last few years, sowed this seed that Apple and Google should play more of a role via the App Store. So you had Antigone Davis, who was Facebook's I was gonna say deceased. She's only, she's left the company. Uh, she's the former global head of

Mike Masnick 20:13

To Zuckerberg, deceased.

Ben Whitelaw 20:16

Yeah. she in 2023 made the point that Apple should do more. Zuckerberg actually, I think, spoke about Apple doing more in a deposition of some sort, um, soon afterwards. And then Yoel Roth, who's been on the podcast and now works at Match Group- Mm in 2024 also said that he would like to see Apple play more of a role. And I think, probably both of those figures were thinking that Apple being a controller of, of the kind of age assurance element was the least worst option, as you say, the, the benevolent dictator, approach. and that they're well-positioned in the stack to do that. As you say, I, we don't know what the potential implications of that are for, children or parents, but it, it, in my mind, it feels like the kinda best of the options right now. But I I obviously can't, I don't have a crystal ball.

Mike Masnick 21:06

and I wouldn't necessarily disagree with that. I mean, of the other options out there of whom I do this, they probably are all worse. And I do think that, like, Apple does a fairly good job of this. But the other side of that is then it also sort of locks Apple in as critical infrastructure in a way that, you know, it was amusing to me to watch Meta really strongly for years they've pushed the idea that, oh, push all of the age verification stuff to Apple and Google so that we don't have to do it. But all that does is really lock in Apple and Google as these central providers who are critical infrastructure and have to be there, and it gives them a lot more power. And, over time, that power is gonna get abused.

Ben Whitelaw 21:48

Yeah. And, and just to be clear, you know, this announcement isn't speaking to the age verification aspect because, we mentioned it on the podcast back in March. that has rolled out in the UK, so in the latest, one of the latest OS updates, you as a UK user had to say that you were above 18, and that information is what gets passed on to app developers. It's not only y- in the UK that that happens, it also happens in Singapore, in South Korea, and I believe in Texas as well. So it's a small number of, states, jurisdictions where that happens. But on, with this new announcement of them, I guess differentiating themselves, let's say, through safety measures like screen time and, the kind of, website approval process, I think there's a really interesting shift there. And, and it is worth noting as well, you know, Apple and Google often get kind of made out to be similar companies, but Google had their own developer conference a few weeks ago and didn't really talk about these kind of safety measures at all. There was a few quite technical, Chrome and, App Store changes, but nothing to this extent. Definitely didn't spend a large chunk of, the presentation talking about trust and safety. And so, I noted that Benedict Evans, who's a kind of industry analyst, noted that this could be Apple's new marketing ploy. in the same way that kind of privacy has been its focus for many, many years, it might be that trust and safety they fear is gonna be a differentiator against Google in a world where parents are very concerned about teen safety and, you know, regulators are gonna be increasingly kind of watching them. So I it's a big signal, I would say, as to how things are moving, and trust and safety professionals probably will be delighted by that. You know, a big company like Apple pinning their, their hopes in some ways on, safety as a, marketing tool, as a, as having good ROI, in kind of industry terms. That will be, that will please a lot of folks who read tech dirt and, and read everything in moderation.

Mike Masnick 23:43

Yeah, and I, and I think it's, it is an outgrowth of their focus on privacy over the last few years. I would argue that they probably see those as, connected. The sort of privacy to trust and safety connection is not, you know, not moving in a different direction. It's sort of leaning further into the idea of, like, protecting the user

Ben Whitelaw 24:00

Mm-hmm. Yep. Agreed. there's a lot more we could say on this. There's I would definitely recommend going to watch that segment again. We will include it in the show notes. There's a couple Apple's trust and safety folks are actually giving, the speech and, up on stage with, the kind of head of engineering who, kind of leads a lot of the, presentation. So it's trust and safety front and center, for sure. We'll move on to our kind of eye-catching section, Mike, where we've got a couple of stories that we wanna, to flag to listeners before we, deep dive into the Anthropic model. we probably should kind of talk a little bit about, before we do that, Canada,

Mike Masnick 24:34

Yes

Ben Whitelaw 24:35

in the last kind of 24 hours has finally got its act together and put out its own digital regulation, kind of update, which, isn't that far away from what the UK has been doing over the last few years

Mike Masnick 24:46

Yeah, I mean, I think it's, part and parcel with what Australia has done and what the UK is doing. Canada is now introducing its own sort of online safety, bill, which, is a little bit vague at this point. they released it just yesterday. a lot of it is sort of left up to, you know, the ideas of setting up regulatory bodies and, which will determine a bunch of the factors. I mean, this is, as far as I can tell, sort of a Canadian specialty where they will sort of release a law and then the actual implementation goes into effect later. It's not true it's just Canada. There's plenty of other countries that do similar sorts of things, as we've seen. but it's clearly this pattern of, setting up rules, limiting children's access to social media, putting more onus on social media companies, including some sort of duty of care, duty of responsibility to try and force the companies to do more to, keep people safe. But, like, the vagueness is part of the problem, where it's easy to say that it's very, very difficult to understand how do you actually do that in a way that doesn't have unintended consequences that, cause more harm than, than they solve. But, you know, like Starmer's speech at the beginning, it's a thing that, on the political front, a lot of folks feel that they need to do this and that they have to focus on these tech companies as being harmful and evil, and therefore there must be a law passed, and this is Canada's attempt to do that. And it, you know, it really picks up on the exact same things that have, come before it in the UK and Australia as well. and we'll, be following the specifics of that as we go, but it's another one of, those kinds of laws, basically.

Ben Whitelaw 26:25

Yeah. And we'll see if, Canada gets a bit of a second mover advantage here because, they've been able to see how the UK and Australia regimes have been rolled out and some of the challenges. So, that will be interesting. talking of Australia, Mike, this is a story that, you know, it makes sense if we've covered two of the big Commonwealth countries to, to also include, a, a third big one. So, Australia announcing that there's gonna be a slight change to how they, regulate some of the companies

Mike Masnick 26:51

Yeah. I mean, this is also, I mean, it, fits with this entire discussion that we've been having today, which is that, these countries are trying to regulate these things. And, and Australia obviously was up front with, their social media ban for teenagers, which as we've discussed over and over again, I would argue has not been particularly successful, and there keeps coming out more and more data showing that, teenagers are still accessing these services. And in fact, the cases where kids have been cut off from the services, there are plenty of stories of people where it was helpful. I mean, there was a whole thing on, children with disabilities who were cut off from support groups because they couldn't access them anymore because of the social media ban. But then you see all the stories of kids tricking or parents helping kids get around these bans and all these sorts of things. So Australia is now, you know, they're sort of framing it as, we're building on this, but, with more enforcement. But what to me that says is, it's an acknowledgment that the social media ban has been a total failure. And so rather than admit that they were going down the wrong road, now they're just going to blame the tech companies for not effectively banning kids from doing what they wanna do, which is talk to their friends and communicate with different communities. And so basically, the plan for Australia now is to ramp up the enforcement side of it. So, basically pulling out the stick just as Starmer was saying, we're gonna start throwing, officials from tech companies into jail. this is Australia sort of figuring out what sort of bigger stick can we take out to force companies to wave the magic wand and suddenly make the internet safe and make sure that no kids ever have anything bad happen to them. and so, to me, it's funny how this particular story is portrayed as, sort of like a continuation of what was already done rather than as an acknowledgment of the complete failure of the, the under-16 ban in Australia, which I think, I think it really is. But it just kinda goes to show that, these things don't stay in one place. And even if you have a particular plan, as soon as you find out, like, kids are still using the internet, bad things are happening to kids because bad things happen to some kids all the time. if you pinned all of the blame on the tech companies, you're just gonna go further and further and try and say, the companies need to magically wave their magic wand and, and stop it

Ben Whitelaw 29:07

and so this, so this kind of is, you mentioned the kind of enforcement regime. There's, there's also an element of kind of a duty of care that, that Australia wants companies to have, and that, is, that trying to get them to kind of be m- kind of more proactive and, preemptive of risk with a focus on, product safety rather than, trying to ban-

Mike Masnick 29:24

Yeah. and that's, always been a, part of a lot of the regulatory approaches. That's in the Canadian one as well. It's, in a bunch of other bills in other places, this idea of a duty of care. you know, the duty of care is one of these things that sounds really good until you really think about it or understand what the impact is. I mean, who could be against a duty of care, right? You want companies to care, right? You want companies to be careful and to think through the consequences of, what they do. But as soon as you put that into the law as a duty, as something that is required, then it becomes a compliance thing. This gets back to Daphne Keller's famous, like, trust and safety becoming a compliance function. And so now suddenly you're checking boxes for a regulatory regime rather than, figuring out how can we experiment and how can we actually do things that make people safer. Instead, it's how can we do things that won't get us held liable, and those are two different things. There may be some overlap, and the approach that people take is different, and now it becomes what is going to limit our liability versus what will actually make people safer. There, obviously is probably some level of overlap, but the reality is that when you have a duty of care, historically what tends to happen is to reduce liability is you just start blocking like crazy. It becomes a kind of censorship regime because you're trying to avoid the liability, and whether or not that makes people safer is a separate question

Ben Whitelaw 30:48

Hmm. Yeah, yeah. Okay. You can quickly see how that could, end badly. I mean, it, it does link actually to our next eye-catching story in a very neat way. I don't know if you did that on purpose. Um, this, this is a, a story from, one of our patrons, Emma. thanks for sending this in, Emma. the kind of original story is about the oversight board, issuing a, case judgment against Meta in relation to a user who, on Instagram sent abuse to a female journalist in which, there was threats of violence, and that user was, banned, permanently. And the case actually kind of, stands, up for Meta and says that that was the right course of action for a n- number of reasons. But what the case judgment goes into is a lot of interesting kind of discussion about how Meta deals with permanent account bans, which has become an increasingly kind of fraught issue. I don't know if... Y- don't have to go very far around the internet, Mike. Comment threads, you probably get, s- people on Tector, there's lots of replies on Twitter, people saying, "I've lost access to my Meta account," mainly. I don't know what to do. It's my small business. It's my, you know, my child's, football account," whatever it might be. I've been summarily banned. I don't know what to do. There are no means of retribution or, or redress." and what the oversight board judgment looks at is the, need for Meta to really improve the way it explains why accounts get banned, the reasons if an account is, taken down and, permanently banned. also the kind of appeals process to be much clearer and much more transparent to allow users to essentially follow their appeal through the system. and again, I think this is, this for me is one of the most underreported stories in lots of ways, is, how many people this is happening to. And we really have no sense of, of it because there's no public data apart from some, reports that are coming out via, out of court dispute settlement bodies in the EU. But it's a real... I feel like we're seeing the kind of, iceberg, the tip of the iceberg when it comes to account takedowns, Mike. and think Emma's point was, there's been some really positive reaction to this judgment from Meta. she flags a, a national coalition against censorship response. A few other, organizations have inputted into the process and have, you know, celebrated what, the oversight board have said. And we, we're yet to see what Meta will do with that. But do you feel also that it's a, as big a problem as, as the oversight board is saying?

Mike Masnick 33:15

as I'm known to do, I, I s- h- I see a lot of nuance in this and trade-offs. So it's like there is an element that, yes, it's always a little worrisome when these companies are total black boxes and you have no idea how the decision-making process gets done and there's no appeals process and there's no clear, way to deal with mistakes, that, customer service is a problem. And as we noted last week, you know, Meta may be trying to improve customer service with AI, but that leads to other kinds of problems. You know, these are, difficult things, and so having more transparency and clear due process and a clear process and for dealing with these things feels good and important. But also the thing that I will remind people of is that you have bad actors who will game those systems, right? So there are reasons why companies, and Meta in particular, I think, tend to be s- fairly secretive about these things, and part of it is that, one, as soon as you're explaining to someone, "This is why we're banning you," if they are a bad actor, if they are nefarious or malicious, they will use that against you, right? First they'll sort of rules lawyer you and say, "Well, I didn't do that. I didn't violate that rule." If you tell them exactly which rule you violated, they will say, "I didn't do that," and they will argue down to this or that point. And often, the really malicious ones will then go to literal court and say, "They said this, and that was defamatory because I didn't break this rule." And it is a, a huge waste of time, and this is why, like, I've seen people say, be really nice to have a, policy for social media that is effectively like the I wasn't born yesterday rule," which is like, "You're just trying to mess with us, and we're banning you because we don't have time for that." and at the scale of Meta, you can't really do that. That is, you know... That doesn't really work, and that's what this ruling, I think, is a kind of response. It's like, no, you actually have to have rules and processes in place, and that includes due process, and that does include transparency. But I think what this ruling and the celebration of it leave out is, not everybody who was banned will accept the sort of transparent explanation or the due process, and in fact, they will just weaponize that in some way or another. And so companies often will sort of be less transparent and have less clear appeals processes because they just, they know that it's just going to be used against them

Ben Whitelaw 35:41

Yeah. I mean, I've spoken about the LinkedIn, issue that I had where

Mike Masnick 35:45

Yes. you were, you were targeted.

Ben Whitelaw 35:47

disappeared, and I had to go to great lengths to, get that page back and up and running. I, I certainly feel like there's a greater degree of information about the process, about where you are in the chain that could help. I agree that actually sharing what's going on behind the scenes could be, problematic in the long run. But again, we get to see what Meta are gonna do with this. they sometimes take the oversight board responses and do something with it. In other cases, they don't, so we'll see. but I think that probably, I'd say, is enough for, for our free listeners this week, Mike. It's probably more than they can take. Um, they also, I think, probably have World Cup games to go and watch and, uh, and listen to. but if you're a sucker for more, this week's Ctrl Alt Speech Plus will look at the rocky start that the Fable 5 model has had and what it means for the way that AI labs think about safety. before we dive in, Mike, why do you think this one deserves a, deeper discussion?

Mike Masnick 36:42

Yeah. I mean, I think obviously, what is happening in the AI world as it relates to online speech and, and online safety is incredibly important, and understanding where that's going and how the big companies are acting and where they're making mistakes is really, really important to understanding more broadly across the internet how speech and safety are going to intersect with one another. And I think that there are a bunch of lessons that can be taken from the way that Anthropic rolled out Fable and Mythos, its, two sort of latest version models this week, and what is maybe working and what is not working. it's a really important discussion to have.

Announcer 37:19

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