WEBVTT
00:00:00.000 --> 00:00:01.627
Mike, I don't know How often you cook,
00:00:02.213 --> 00:00:02.592
Oh.
00:00:04.003 --> 00:00:08.243
But you will have known by now that following a recipe on your phone is an absolute nightmare.
00:00:08.369 --> 00:00:18.414
I've got so bored of like, constant scroll and the skipping between videos and, like, those kind of SEO bait recipes that have become a bane of all of our lives.
00:00:18.672 --> 00:00:22.853
Well, I found a solution, and it's very, very cool.
00:00:22.853 --> 00:00:31.527
It's this thing called Recipe Tables that rather than showing a recipe as a story, has come up with, with what it calls a, prep table.
00:00:31.647 --> 00:00:42.450
And so recipes are now shown visually as a table with the ingredients down one side and the techniques that you need to use to make your meal on the right-hand side.
00:00:42.450 --> 00:00:45.200
So you can see at a glance what you need to do and when.
00:00:45.210 --> 00:00:46.670
You don't have to go back and
00:00:47.012 --> 00:00:47.893
No scrolling.
00:00:48.119 --> 00:00:49.060
it's no scrolling.
00:00:49.079 --> 00:00:49.939
It's fantastic.
00:00:50.390 --> 00:00:59.582
Um, and this tool allows you to paste any URL into its tool, and it kind of displays your recipe, exactly in this kind of special prep table form.
00:00:59.593 --> 00:01:00.383
So it's wicked.
00:01:00.482 --> 00:01:01.103
It's really cool.
00:01:01.442 --> 00:01:01.923
Very nice.
00:01:01.923 --> 00:01:05.763
It's very old-fashioned though, because the kids these days are using AI to help them.
00:01:06.272 --> 00:01:10.718
Just ask the AI to tell them what other ingredients they need for their recipe.
00:01:10.992 --> 00:01:11.263
True.
00:01:11.263 --> 00:01:11.783
Yeah, true.
00:01:11.783 --> 00:01:13.162
What have I, what steps have I missed?
00:01:13.462 --> 00:01:13.492
Yes
00:01:14.643 --> 00:01:16.069
what have you been looking at this week?
00:01:16.411 --> 00:01:21.355
Uh, so the, the really cool thing that I saw on the internet that reminded me of how cool the internet can be...
00:01:21.408 --> 00:01:23.207
do you know who Rainbolt is?
00:01:23.847 --> 00:01:25.378
There's a YouTuber named Rainbolt.
00:01:25.408 --> 00:01:26.418
Have you ever heard of him?
00:01:27.048 --> 00:01:32.918
So his sort of claim to fame is that he is very, very good at the game GeoGuessr.
00:01:32.957 --> 00:01:34.287
Have you ever played GeoGuessr?
00:01:34.477 --> 00:01:34.867
Yes.
00:01:34.878 --> 00:01:36.275
So very fun, whatever.
00:01:36.275 --> 00:01:43.564
So Rainbolt is sort of a, a YouTube influencer who is very good, though my children tell me he is not the best at GeoGuessr.
00:01:43.564 --> 00:01:47.834
He is merely a good YouTuber at GeoGuessr.
00:01:48.051 --> 00:01:54.840
but occasionally, you know, he'll, he'll do other kinds of videos, but most of them around his sort of skills at finding stuff.
00:01:55.350 --> 00:02:05.775
So he put up this video that I warn you ahead of time may make you cry, because someone had reached out to him.
00:02:05.795 --> 00:02:08.745
'Cause, like, people will ask him,"Can you find where this was?" They'll like...
00:02:09.025 --> 00:02:18.594
I've seen him find, like, photos of, like, great-grandparents on their honeymoon in some random place, and he'll figure out where it was, you know, 120 years ago or whatever.
00:02:18.935 --> 00:02:30.764
But this one was, someone said that when they were a kid, they were, like, seven or eight years old or something, and their father had taken them to a Lakers game, basketball, back to
00:02:30.919 --> 00:02:31.000
Ja.
00:02:31.578 --> 00:02:35.617
and that the father had unexpectedly passed away soon afterwards.
00:02:36.057 --> 00:02:42.777
And they, they had a photo from the game, but the kid did not know which game it was and was trying to find which game it was.
00:02:42.777 --> 00:03:05.842
So he gave him the photo and he said,"I promised I would find it." And then he systematically goes through how he used the internet and, like, Getty Images and, news stories, and all of these other, like, little things, and eventually, I'm not gonna, away exactly how he finds it, but at one point he's like, The real secret to all of this were these two blue pixels." And you're just like, oh my God.
00:03:06.342 --> 00:03:08.293
And he figures out which game it is.
00:03:08.383 --> 00:03:30.062
Then he finds online the video of the game, watches the entire video of the game until he spots them, the kid and his father, in the audience, and it's just like, it's, it's really, very, very moving, both incredible about the internet and what can be found on the internet and how it can do it, and then just a very, very moving, ending for that.
00:03:30.112 --> 00:03:36.543
And if you look also, I will say, read the first comment on that video after you've watched the video as well.
00:03:36.562 --> 00:03:36.872
So
00:03:37.187 --> 00:03:37.687
Okay.
00:03:37.796 --> 00:03:39.616
Little Easter egg that you've left us there.
00:03:39.776 --> 00:03:41.977
I mean, that s- that sounds fantastic.
00:03:42.067 --> 00:03:50.759
It's, it's the perfect tonic it sounds like to this, today's episode, where we're gonna, we're gonna be talking about, nothing so wholesome
00:03:51.138 --> 00:03:51.747
Yes.
00:03:52.800 --> 00:03:53.539
unfortunately.
00:04:02.033 --> 00:04:11.122
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.
00:04:11.436 --> 00:04:25.391
It's August 27th, 2026, and this week we're going all in on the big story of the week, Meta's blockbuster online safety settlement, what it entails, what they've agreed, and for our extended Patreon episode, why it matters and what happens next.
00:04:25.627 --> 00:04:38.291
Ctrl Alt Speech is the podcast that, unlike Meta, settles almost nothing, l- leaves almost everything open, and it's co-hosted by Mike Masnick, founder and editor of Techdirt, and me, Ben Whitelaw, founder and editor of Everything in Moderation.
00:04:38.732 --> 00:04:41.278
Mike, this podcast doesn't cost 17 billion pounds.
00:04:41.447 --> 00:04:41.708
Uh,
00:04:42.057 --> 00:04:42.627
Not yet.
00:04:44.108 --> 00:04:45.088
Give us time, Ben.
00:04:45.088 --> 00:04:45.767
Let us cook.
00:04:47.137 --> 00:04:49.305
it's, it's a slow and, gradual growth.
00:04:49.483 --> 00:04:51.682
but yeah, this, this has been the massive story of the week, right?
00:04:51.682 --> 00:04:52.372
It's, it's...
00:04:52.622 --> 00:05:00.485
I like it when there are stories that, we care about that get covered in the mainstream news, and which kind of people will, almost stop you in the street to talk about, right?
00:05:00.485 --> 00:05:01.675
This, this has been one of those.
00:05:01.896 --> 00:05:02.425
Yeah.
00:05:02.475 --> 00:05:11.466
Though, it, I mean, it's, it's crazy because it came out yesterday and, all week long between last week's episode, which was a very good episode you missed, but was still a very good episode.
00:05:11.651 --> 00:05:19.658
and there were a bunch of stories, and I was filling up our, spreadsheet with all the different stories we wanted to cover, and then yesterday they all got wiped out
00:05:20.490 --> 00:05:20.651
Yeah.
00:05:21.379 --> 00:05:23.569
'cause we've got a story to cover today.
00:05:24.019 --> 00:05:24.428
Yeah.
00:05:24.488 --> 00:05:29.369
It was a giant meteor of a story that landed on the Ctrl Alt-Speech Earth.
00:05:29.536 --> 00:05:35.178
we can't really not talk about anything else, and there's so much to it that w- we, couldn't really, even cover it in just a free episode.
00:05:35.178 --> 00:05:38.899
I think we're gonna have to split it out, really break it down, talk through what it looks like.
00:05:39.206 --> 00:05:40.408
we'll get into it basically.
00:05:40.509 --> 00:05:44.569
before we do, you know, the people who haven't settled are our Patreon supporters, Mike.
00:05:44.829 --> 00:05:49.716
They, they are, investing in, in a way that is almost akin to meta
00:05:49.740 --> 00:05:52.610
can we get attorneys general to force people to pay us?
00:05:52.610 --> 00:05:54.029
Is that, is that where we're getting at?
00:05:54.841 --> 00:05:58.346
I think this is a good, this, settlement suggests we might have a case.
00:05:58.874 --> 00:05:59.454
There we go.
00:05:59.836 --> 00:06:00.276
Yeah.
00:06:00.386 --> 00:06:03.326
Um, but we've got, uh, we've got lots of Patreon supporters.
00:06:03.326 --> 00:06:05.685
We've got our free supporters, who we love.
00:06:05.685 --> 00:06:08.286
They're dipping their toe into the waters of Ctrl Alt-Speech.
00:06:08.565 --> 00:06:11.331
they're not paying anything right now, but they're getting a sense of what we provide.
00:06:11.661 --> 00:06:21.262
We've got our supporter tier that are getting our extended episodes and earlier access for, much less than£18 billion, and that is$10 a month.
00:06:21.721 --> 00:06:27.982
And we've got our founder and insider tiers who are getting all that and the chance to suggest stories for$20 a month.
00:06:28.281 --> 00:06:37.478
And as of today, we are rolling out the chance for those supporters to give us ideas about what special episodes we should be doing for our Patreon supporters.
00:06:37.478 --> 00:06:47.336
So if you are an insider or a founder, supporter and you have something you want us to cover, a particular question you have, we are taking ideas for special episodes of Ctrl Alt-Speech.
00:06:47.646 --> 00:06:51.615
We've already done one, which is our s- our reading list episode, which we put out a couple of months
00:06:51.886 --> 00:06:53.475
Which was really good, by the way.
00:06:53.505 --> 00:07:03.935
If, if you haven't listened to it, it was, you know, we each went through three different books that we think are really incredibly valuable if you're interested in the kinds of topics that we talk about.
00:07:04.233 --> 00:07:06.769
and we got some really good feedback on this as well.
00:07:06.809 --> 00:07:14.718
We may end up doing another episode with more books because as we noted, there are a lot of very interesting books in the space, but that was a really, really fun episode
00:07:15.170 --> 00:07:15.670
Definitely.
00:07:15.880 --> 00:07:21.610
I've been questioning whether I included the right three books in that ep- episode, so we will have to do another one.
00:07:21.697 --> 00:07:26.278
but yeah, our Patreon supporters are a massive part of how we produce and sustain Ctrl Alt Speech.
00:07:26.401 --> 00:07:32.074
we also, have some brilliant sponsors who come on board to sponsor and, and reach our excellent audience.
00:07:32.274 --> 00:07:37.913
So if you're interested in becoming a supporter, go to patreon.com/ctrlaltspeech, more information there.
00:07:38.983 --> 00:07:47.141
And if you want to sponsor us, go to the Ctrl Alt Speech website and you can find out more information and some examples of, sponsorships that we've done in the past.
00:07:47.420 --> 00:07:55.197
as we get to the end of August, Mike, as, as August comes to a close, it's important to mention that our Trust Con code is still valid for
00:07:56.093 --> 00:07:57.504
Our discount code, yes
00:07:57.627 --> 00:08:05.468
Our discount code, If you're interested in, doing more than dipping your toe into the waters of Ctrl Alt Speech and you want 90% off your first month, use Trust Con.
00:08:05.870 --> 00:08:10.774
we're there at the end of August doing our live episode, and this is a special promotion we're running.
00:08:11.033 --> 00:08:15.468
that's still available until the end of this month, so ending on Tuesday.
00:08:15.935 --> 00:08:16.485
Yes
00:08:16.769 --> 00:08:17.230
Perfect.
00:08:18.250 --> 00:08:19.610
Glad, glad we're agreed on that.
00:08:20.588 --> 00:08:20.997
No.
00:08:21.548 --> 00:08:22.218
How dare you?
00:08:22.718 --> 00:08:29.447
Uh, yeah but, uh, I mean, we put our all into this podcast, and, we hope that that comes through in what you hear.
00:08:29.747 --> 00:08:45.914
and if you wanna see it keep going and continue to support the work that we're doing in trying to explain these complex topics and make them understandable and make them real, please look into supporting us on, Patreon, and thank you already to those who have already stepped up and support us there.
00:08:46.403 --> 00:08:52.235
go subscribe, and you can do it at a discounted rate if you use the code TrustCon for the next few days
00:08:52.658 --> 00:08:52.948
Indeed.
00:08:53.239 --> 00:09:00.418
And if, if you, want to become a paid supporter right now, you can obviously help us out by rating and reviewing us wherever you get your podcasts.
00:09:00.428 --> 00:09:01.158
That is free.
00:09:01.339 --> 00:09:04.711
all it needs is your time and your, creativity, let's say
00:09:05.148 --> 00:09:08.357
And also, you can promote the podcast to other people.
00:09:08.418 --> 00:09:18.577
You can tell other people about the podcast or about the stories that you liked or that you, we said something that you found interesting or compelling or that changed your mind on something.
00:09:18.618 --> 00:09:19.687
Tell other people about it.
00:09:19.738 --> 00:09:21.717
Word of mouth is how podcasts grow.
00:09:21.758 --> 00:09:23.827
So we always say rate, review, and subscribe.
00:09:23.827 --> 00:09:30.004
That is the sort of official ways, but I think we should be more aggressive in reminding people to tell other people.
00:09:30.004 --> 00:09:35.162
People you know may be interested in what we're saying, and the way they're gonna find out about it is if you tell them
00:09:35.566 --> 00:09:35.936
Yeah.
00:09:36.115 --> 00:09:43.958
Ping that Discord group or that Slack channel or, you know, those friends that you don't really like but think they should have better views on the internet.
00:09:44.219 --> 00:09:46.408
you're the smart person who knows about the internet.
00:09:46.778 --> 00:09:50.489
Do your thing and, uh, share the good word of Ctrl Alt Speech.
00:09:50.918 --> 00:09:55.042
Um, Mike, we really have a lot to get into in so far as this Meta settlement goes.
00:09:55.042 --> 00:09:58.645
This is, a case that's been rumbling since 2023.
00:09:58.895 --> 00:10:01.812
It might have caught a lot of people off guard if they've not been paying attention.
00:10:01.985 --> 00:10:05.525
and so you've written about it extensively this week on Techdirt.
00:10:05.812 --> 00:10:10.408
I've seen a lot of people sharing your post, as a, counterpoint to a lot of the reporting out there.
00:10:10.708 --> 00:10:19.609
But just before we dive into some of those ideas, talk us through what the actual settlement entails, and some of the kind of things that have been agreed off the back of it.
00:10:20.144 --> 00:10:20.514
Yeah.
00:10:20.514 --> 00:10:40.738
So this, this was a case that was bo- brought by most state attorneys general in the US, not all, most of them, against Meta a few years ago, you know, claiming all of the usual things about child safety and, and, you know, the argument being that Meta does not take child safety seriously enough and something needs to be done.
00:10:41.039 --> 00:10:45.238
And the case has been moving forward in sort of fits and starts, over the last few years.
00:10:45.347 --> 00:10:51.727
And, there's the trial that's going on in California right now that was sort of related to that.
00:10:51.817 --> 00:10:59.195
there's obviously also the New Mexico case that we covered, a few months ago when, when all of that happened and, and the rulings there.
00:10:59.475 --> 00:11:10.072
but what happened yesterday as we're recording was that Meta settled with 52 attorneys general, and basically wiped out most of those cases.
00:11:10.317 --> 00:11:11.725
you know, it was a consolidated case.
00:11:11.745 --> 00:11:32.490
But, wiped out that, made an agreement they were gonna pay a bunch of money, that the amount that they pay depends on a few different factors, which we'll get into in a, a moment, they have promised to put in place a bunch of different changes within the app, which are designated to hopefully protect and help protect children in some form or another.
00:11:32.788 --> 00:11:37.807
and they are also going to try to make this into an industry standard.
00:11:37.817 --> 00:11:54.528
The features and changes that they're making, or at least some of the features and changes that they're making, they are trying to impress upon specifically YouTube and TikTok to make the same changes, and there are some incentive structures involved here to get them to try and make it an industry standard.
00:11:54.567 --> 00:12:07.548
And if other social media companies reach to a certain level, then Meta is also effectively required to try to pressure them into implementing the same features as well.
00:12:07.825 --> 00:12:09.404
so that's the sort of headline.
00:12:09.754 --> 00:12:16.528
You know, the headline number that gets a lot of attention is that they're paying, it depends on which reporting you see.
00:12:16.538 --> 00:12:19.418
It's somewhere between$16 and$18 billion.
00:12:20.488 --> 00:12:22.687
and that money will go to the states.
00:12:22.697 --> 00:12:28.087
In theory, that money is supposed to be used for child safety related things.
00:12:28.501 --> 00:12:31.572
it will be paid out over the course of about 10 years.
00:12:31.822 --> 00:12:39.261
you know, so a lot of people, the initial response is Meta is a very large company that makes a lot of money.
00:12:39.534 --> 00:12:43.725
$18 billion over 10 years is not that much
00:12:44.099 --> 00:12:50.947
and actually to put that into context, revenue last year from Meta's various businesses and apps was 200 billion.
00:12:51.087 --> 00:12:54.134
So we're talking about, less than a 10th of last year's revenue
00:12:54.490 --> 00:12:55.809
Spread out over 10 years
00:12:55.957 --> 00:12:58.227
spread, spread out over 10 years, which we, which we talk about.
00:12:58.600 --> 00:13:04.288
they had a particularly good last quarter, I think something like 24 or 25% up year on year.
00:13:04.298 --> 00:13:08.298
So they made 60 billion just in the last quarter alone.
00:13:08.508 --> 00:13:12.260
So it's essentially one third of last quarter's revenue.
00:13:12.691 --> 00:13:17.563
So we're talking about a month's worth of Meta revenue at the current trajectory.
00:13:17.864 --> 00:13:19.043
That's a really important thing to note.
00:13:19.114 --> 00:13:24.114
So in salary terms, you know, I was thinking about it in terms of like my, my salary take home.
00:13:24.254 --> 00:13:26.624
It's a bit like having to replace your boiler,
00:13:27.974 --> 00:13:28.384
Yes.
00:13:28.604 --> 00:13:28.933
right?
00:13:28.964 --> 00:13:29.833
You know, for, for your
00:13:29.903 --> 00:13:30.293
that's a...
00:13:30.403 --> 00:13:35.947
your average person, you know, again, you know, not, not giving away too much about how much I earn here, I don't think.
00:13:35.957 --> 00:13:39.378
But like boiler, bo- boilers vary in, in, in cost.
00:13:39.638 --> 00:13:43.980
But for the average person, it's probably the cost of, an unexpected boiler replacement
00:13:44.418 --> 00:13:44.818
Right.
00:13:45.019 --> 00:13:58.875
you know, I, I think I, I sort of said, it probably stings, but you, it's the kind of sting that you forget about a day later, it is not nothing, but it is something that, effectively will be forgotten by the company very soon
00:13:59.337 --> 00:13:59.508
Yeah.
00:13:59.597 --> 00:14:02.467
And in terms of the usefulness, I mean, boilers are useful.
00:14:02.628 --> 00:14:06.394
We'll talk about how useful this settlement will be for Meta going forward.
00:14:06.394 --> 00:14:08.533
But I think it's important to kind of contextualize that number.
00:14:08.933 --> 00:14:14.870
Talk about the 10 years, Mike,'cause I had to do a bit of, uh, I'm not familiar with a lot of, like, US legal settlements.
00:14:15.250 --> 00:14:17.860
The 10 years kind of matters for a few reasons, right?
00:14:18.288 --> 00:14:22.768
the reason it's spread out, as far as I'm understanding, is that they don't they can't just pay it and forget about it.
00:14:23.138 --> 00:14:23.847
Is that part of it?
00:14:24.500 --> 00:14:27.399
I mean, there's a bunch of moving parts within the settlement.
00:14:27.462 --> 00:14:31.525
and it's not entirely uncommon that you would have this kind of settlement.
00:14:31.596 --> 00:14:35.596
you know, where, where things are paid out over time, where things need to be checked over time.
00:14:35.596 --> 00:14:42.515
There's like, there's a process in which an independent auditor has to get in-involved to make sure that Meta is taking care of these things.
00:14:42.515 --> 00:14:49.296
There are other structures that are being set up as a part of this as well, which is what some of the, the money will get paid into.
00:14:49.655 --> 00:14:53.169
but was there something specific about the 10-year structure that, struck you as
00:14:53.312 --> 00:14:57.231
Well, just as a kind of fairly arbitrary number, um, right?
00:14:57.261 --> 00:14:59.655
You know, it's kind of, I was thinking like why 10 years?
00:14:59.676 --> 00:15:00.591
Why not, five?
00:15:00.591 --> 00:15:02.528
Or why not kind of fewer than, than that?
00:15:02.528 --> 00:15:05.479
But, is it fairly common that these kind of settlements would be over a
00:15:05.899 --> 00:15:12.374
Yeah, this, doesn't strike me-- I mean, you know, this is obviously unique in certain ways, but it's not uncommon for...
00:15:12.433 --> 00:15:26.086
You know, the thing that I would compare the ten-year structure to is the various FTC, consent decrees that basically all of the social media companies ha-have had whenever they've done something bad on the privacy front.
00:15:26.086 --> 00:15:34.466
At some point, they come to a settlement decree with, with the FTC, and those usually stretch over a period of time, which will be ten to twenty years.
00:15:34.787 --> 00:16:00.769
And part of the thinking there is generally like, you know, we wanna make sure that you are continuing to do the things that you are required to do over this time, so we're going to be checking back in, and we're putting in place some rules on, on how you do it, which, you know, there is related to the payments over ten years, the, features that Meta has agreed to, make standard or default, those are required over a period of five to ten years as well.
00:16:00.769 --> 00:16:11.167
So it's, it's sort of stretching out this, this sense of like, know, Meta is committed to doing this and can't roll it back or change things very quickly, after the settlement has been agreed to.
00:16:11.490 --> 00:16:16.303
I don't know that the, payment necessarily directly matters on that front.
00:16:16.303 --> 00:16:22.509
I do think that, you know, it also in some sense keeps Meta engaged, that they don't forget about this.
00:16:22.539 --> 00:16:24.600
Like, it, it almost acts as a reminder.
00:16:25.024 --> 00:16:36.149
on the flip side too, it's like if you're the states and you're getting this settlement, whereas it may be a sort of a drop in the bucket for Meta, it's still a significant amount of money that is suddenly gonna go into the state coffers.
00:16:36.399 --> 00:16:43.500
And if it was all a one-time payment, you know, there may be some benefits to doing it that way, but it's also the kind of thing that easily could then get squandered.
00:16:43.750 --> 00:16:56.850
Whereas if you know that you have a ten-year period where you're gonna get, uh, hundred million dollars a year or whatever, you can sort of plan for that and put that to good use with a sort of expectation of, what's gonna happen over the next ten years.
00:16:56.947 --> 00:17:01.376
and so I think the hope is that if you do it that way, the money will actually put to good use.
00:17:01.543 --> 00:17:08.304
you know, we'll see how the states eventually actually use it or how much transparency it-- there is into how the states use it right now.
00:17:08.574 --> 00:17:10.854
I don't know that there is that much transparency.
00:17:11.166 --> 00:17:12.876
I mean, I think that would be really interesting.
00:17:12.886 --> 00:17:18.396
Are the states actually going to put this into child safety in any meaningful way, and what does that mean?
00:17:18.406 --> 00:17:19.227
What does that look like?
00:17:19.227 --> 00:17:21.527
Are the states in agreement of how they're gonna use the money?
00:17:21.876 --> 00:17:26.420
That's information that as far as I know, haven't been told yet, and I don't know if we'll ever be told.
00:17:26.794 --> 00:17:26.993
Right.
00:17:27.213 --> 00:17:27.534
Yeah.
00:17:27.564 --> 00:17:28.034
Interesting.
00:17:28.054 --> 00:17:28.344
Okay.
00:17:28.624 --> 00:17:32.557
so that covers off the amount and the kind of time span that this settlement is over.
00:17:32.938 --> 00:17:49.153
Talk us through the kind of what Meta has agreed to, um, as part of this then, because, yeah, I, I read in your tech, uh, post that you felt it, to some degree that what they've agreed to in terms of product changes actually doesn't really matter, but it's still what's being reported fairly heavily.
00:17:49.412 --> 00:17:49.832
Yeah.
00:17:49.862 --> 00:18:00.271
I mean, it matters to users of Meta's b- uh, products, especially the teen users, because some of the changes are things that aren't really changes.
00:18:00.294 --> 00:18:08.526
A lot of them are things that Meta was kind of already doing or already embracing or already pushed into by other legal processes over, over the last few years.
00:18:08.892 --> 00:18:29.419
so there are a bunch of things, and they're all sort of features that, if you read them, you might say, like,"Oh, okay, these are things designed to, protect teenagers." the big one on, my side is the, age assurance, the guarantee of doing age assurance, which age assurance is the polite way of saying age verification, which is the polite way of saying identity verification.
00:18:29.788 --> 00:18:44.885
age assurance, they always sort of hand wave around it where it's like,"Oh, if we're doing age assurance, we don't necessarily need to get your ID." But generally these things lead back to you have to provide some sort of identifi- the service in some form or another.
00:18:45.281 --> 00:18:55.827
I suppose there's a lot of, uh, margin for error in age assurance, which kind of looks at a biometric image or some other kind of signals, right?
00:18:55.877 --> 00:18:56.288
Yeah.
00:18:56.386 --> 00:19:16.028
some of the issue is just, like, if you're having, like, hard cliffs within your age categorization system, if it's like 16 and under has to be treated this way, the general age assurance is not going to be that good in that the age assurance system might be able to tell the difference between a 12-year-old and a 60-year-old.
00:19:16.382 --> 00:19:18.929
Anyone can tell the difference between a 12-year-old and a 60-year-old.
00:19:19.199 --> 00:19:36.067
Where it gets very tricky is when you're trying to tell the difference between, like, a 14-year-old and a 17-year-old, and if that is important to the systems that you're putting in place, and that can matter because the types of things, experiences, and content that are appropriate for a 14-year-old could be very different than for a 17-year-old.
00:19:36.163 --> 00:19:54.663
and so if you're trying to judge down to that very granular level of, like, between 15 and 17, you generally will need some sort of identification to really prove who someone is, and then you're getting into the question of, like, privacy and the rights of teenagers.
00:19:54.693 --> 00:20:01.818
And as we discussed last week when you were away, France just threw out their, social media ban.
00:20:01.847 --> 00:20:11.487
the constitutional council there threw out the ban that, Macron put in place because they said, like, it is ignoring the fact that teenagers have rights, too, and they do.
00:20:11.608 --> 00:20:21.573
you know, and I've said in the past, like, when platforms are deciding to put in place age-appropriate systems on their own, that is perhaps more understandable.
00:20:21.573 --> 00:20:33.019
But when it's, at the behest of the state, then it begins to get tricky and begins to tiptoe into areas where it's like now the government is trampling on the rights specifically of teens.
00:20:33.257 --> 00:20:45.756
The other aspect of it is that when you're in the age verification, identity verification world, it's easy to say,"Well, you know, we have to protect the kids," but the only way that works is if you identify everybody, right?
00:20:45.760 --> 00:21:03.201
If you're gonna have to know who the kids are, who the teens are, you also have to know who the adults are, which means you're basically doing some sort of process to identify everyone, and then that becomes a sort of big privacy concern over how that is done and how you protect that data and how you keep people safe.
00:21:03.579 --> 00:21:03.890
Yeah.
00:21:03.950 --> 00:21:17.882
And it, and it's worth noting that Meta has in recent months talked about how it is expanding its age assurance measures and using AI to, for example, place teens into its kind of what it calls teen-appropriate experiences.
00:21:17.892 --> 00:21:19.686
So it has been doing elements of this.
00:21:19.894 --> 00:21:23.396
this settlement kind of, I guess, makes that a blanket, thing.
00:21:23.631 --> 00:21:30.300
again, we don't think we know exactly what means they're using, but it's an experience that's gonna be coming to more, more users on Meta.
00:21:30.795 --> 00:21:33.255
so that is one aspect, but there are other aspects as well, right?
00:21:33.255 --> 00:21:40.883
There are lots of kind of little elements that have been given to people who, who are concerned about children's safety
00:21:41.186 --> 00:21:44.386
a whole bunch of other little tweaks and changes.
00:21:44.707 --> 00:21:49.017
The one that I think has probably gotten the biggest headlines is the two-hour time limit.
00:21:49.446 --> 00:21:51.257
So this is across meta platforms.
00:21:51.257 --> 00:21:57.369
Teenagers will be limited to two hours of time per day, and then they'll be cut off.
00:21:57.394 --> 00:21:58.960
there are also regular prompts.
00:21:58.960 --> 00:22:05.173
If you're, like, looking at content for a while, it will remind you you've been looking at this for this much time, sort of leading up to the two hours.
00:22:05.532 --> 00:22:08.252
Now, there are a couple important things on that two hours.
00:22:08.474 --> 00:22:13.587
one of which, which I think is a really big deal, which is that parents can override that.
00:22:14.028 --> 00:22:20.477
and I think there are going to be many stories coming out of, you know, kids are going to be using it.
00:22:20.477 --> 00:22:21.416
They're gonna hit the two hours.
00:22:21.416 --> 00:22:24.836
They're gonna run to the parent and say, like,"Hey, you know, I was in the middle of something.
00:22:24.836 --> 00:22:32.255
I need to get a pass this." And a bunch of parents are going to end up removing that, and then it's not gonna matter, right?
00:22:32.255 --> 00:22:33.845
And so, like, you have all sorts of things.
00:22:33.845 --> 00:22:44.565
There's always issues of, like, when you need parental consent, that raises a whole other, series of concerns about relationships between children and their parents and who has authority and, and whether or not...
00:22:44.964 --> 00:22:55.583
You know, it's great if kids have a good relationship with their parents, and I think it's important for, you know, kids to have parents that they can, communicate with, but not everybody is, so fortunate.
00:22:55.803 --> 00:23:01.911
and you could see where that could lead to conflict and real problems and, issues.
00:23:02.188 --> 00:23:16.932
you know, it has these other features like, it will remove showing you likes because there's this feeling that, like, people start doing things just for the likes, which, is anecdotally very compelling, but I'm not sure that the research actually supports that.
00:23:17.281 --> 00:23:26.190
you know, there's things about, being able to have a non-algorithmic feed, which is something that a lot of people claim is really important.
00:23:26.200 --> 00:23:29.069
But again, the research doesn't necessarily support that.
00:23:29.069 --> 00:23:39.567
In fact, the research has shown that if you, force people into a non-algorithmic feed, they tend to see worse stuff because part of the algorithm is often, finding the better content.
00:23:39.846 --> 00:23:44.896
I will note, going back to my intro, that Rainbolt video was recommended to me by the YouTube algorithm.
00:23:44.926 --> 00:23:45.757
That's how I found it.
00:23:45.876 --> 00:23:48.646
So sometimes the algorithm recommends good stuff.
00:23:49.067 --> 00:23:54.813
Um, so there's, there's a bunch of those kinds of features that, like, people have raised concerns about.
00:23:55.047 --> 00:24:05.192
Where another area, another feature that I think or feature th-thing that I think is concerning is they have, like, uh, specific categories of content.
00:24:05.192 --> 00:24:10.442
So the specific one is disabling cosmetic surgery and extreme makeup filters.
00:24:10.878 --> 00:24:48.182
and they define it fairly narrowly and carefully, and they say, like,"Oh, it doesn't include these other things that, aren't problematic." And, you know, you can look at this and say, Okay, I can understand where there are concerns about, encouraging teenagers to, do drastic cosmetic surgery or things like that." But we're getting into spaces where this is constitutionally protected speech, and as soon as you're in this space where, like, the government has pressured a company into saying, like,"We have to block this specifically defined content from children," it becomes really, really problematic.
00:24:48.553 --> 00:25:00.348
And the issue that I raised in my piece on it is, a lot of these things, like, if you just look at them, y- you say, like,"Sure, like, maybe kids shouldn't have access to this." And I, and I understand why Meta's doing them.
00:25:00.598 --> 00:25:05.740
But many of the feature changes, we don't know if they actually protect children.
00:25:06.144 --> 00:25:13.635
and you can make cases where they probably protect some children, but you can also look at this, this stuff, and I, I've been having discussions...
00:25:13.885 --> 00:25:54.313
Actually, just this morning before we were even recording, I was having a discussion with someone who, who has some expertise on, suicide prevention, and was looking at some of the feature changes and saying, This might be really damaging to kids going through some mental health issues, where if these are not implemented well, it could cause, teenagers dealing with a mental health episode to spiral in some ways." and so there is this element of, people talk about, like, oh, it's unfair or it's bad that companies are experimenting on our children with their algorithmic feeds or the different features and, you know, we don't know if it's safe or whatever.
00:25:54.593 --> 00:26:11.055
But this is also an experiment, and this is because in some back room somewhere, a bunch of Meta lawyers and a bunch of state attorneys general got together and just decided in that room what they think will make kids safer.
00:26:11.681 --> 00:26:23.030
And I don't think that there were probably that many child safety experts in that room, and we don't actually know if any of these feature changes will actually protect kids.
00:26:23.050 --> 00:26:25.270
Will, in some cases, will it make them worse?
00:26:25.760 --> 00:26:27.711
And there doesn't seem to be...
00:26:27.740 --> 00:26:35.768
You know, the settlement agreement does allow Meta to study some of these issues, but, you know, they're still sort of locked into them.
00:26:35.998 --> 00:26:46.425
And there is, like, an independent auditor that is a part of the process, but the independent auditor is to, audit that Meta is actually doing these things, not that they're working or protecting children in any way.
00:26:46.607 --> 00:26:46.968
Yeah.
00:26:46.968 --> 00:27:06.036
Well, there is this, independent social media research foundation, which it sounds nice, but there is not a lot of detail about what information would be available to researchers, academics, whether that would be historical data, live data, and whether it would kind of chronicle, this period under which these changes will be made.
00:27:06.256 --> 00:27:07.801
So I think that, that's a fair call out.
00:27:07.801 --> 00:27:14.641
I, I also wanted to kind of point out, I guess, on the time limits, two hours a day across Meta products sounds good.
00:27:14.921 --> 00:27:21.570
I think we, we had this idea that, teenagers are using their phones a lot on social media, that that maybe is a reduction.
00:27:21.830 --> 00:27:32.084
I was looking up, one poll, one Gallup poll in advance of chatting, and, and there's apparently boys and, girls between the age of 13 and 18 spend about five hours on social media.
00:27:32.395 --> 00:27:35.997
And Meta are not the top most used
00:27:36.230 --> 00:27:39.611
I was gonna say it's not Meta that they're spending that time on.
00:27:39.717 --> 00:27:40.186
right?
00:27:40.317 --> 00:27:47.027
So, so according to this one poll based upon 2,000 US teens, YouTube is, is around two hours.
00:27:47.836 --> 00:27:51.999
TikTok is, around, an hour and a half across boys and, and girls.
00:27:52.138 --> 00:27:53.269
Instagram is an hour.
00:27:53.298 --> 00:27:55.481
Facebook is, 3 of a, of an hour.
00:27:55.501 --> 00:28:08.830
So and it will come onto this, and I think it's relevant because we- we're gonna talk about Meta becoming an industry standard and how they are, they've made some contingencies within the settlement, that relate to their competitors, YouTube and TikTok.
00:28:08.830 --> 00:28:11.451
So I think this two hours sounds like a, a lot.
00:28:11.809 --> 00:28:14.558
it maybe isn't as, as little as we thought
00:28:15.295 --> 00:28:15.595
Yeah.
00:28:15.625 --> 00:28:30.863
I mean, when you look at it, and then we'll get into what the requirements are for other platforms, but yeah, somebody pointed out, like, this would not actually probably diminish that much usage of Meta products because most people don't use them more than two hours.
00:28:30.863 --> 00:28:33.212
It is the combined with other platforms.
00:28:33.722 --> 00:28:42.730
But, you know, if you look at the other platforms, the YouTube and the TikTok, a lot of that interaction and engagement is a different type, right?
00:28:42.881 --> 00:28:46.451
YouTube is more sort of sit back and watch engagement.
00:28:46.461 --> 00:28:55.048
That can be true of, Instagram, but Instagram is l- usually a little bit more engaged and involved and, reading, replying, all that kind of stuff.
00:28:55.348 --> 00:28:57.288
TikTok, I think, is somewhere in between.
00:28:57.338 --> 00:29:02.419
You know, it is a lot of just sort of watching videos, but also some level of engagement.
00:29:02.852 --> 00:29:06.211
and I think YouTube is-- tends to be more just sit back and watch.
00:29:06.221 --> 00:29:12.508
More people use YouTube as a kind of modern TV rather than social media in that sense.
00:29:12.728 --> 00:29:21.266
And so, this gets at the thing that, like, all these different platforms are somewhat different, and treating them all as the same can be problematic.
00:29:21.556 --> 00:29:40.434
Which gets to the next concern of this agreement and the weirdest part of the agreement, which is that Meta has to try to convince specifically YouTube and TikTok to put in place the same features, or not all of the same features, but most of the same features, and limits.
00:29:40.684 --> 00:29:44.515
And that is weird on almost every level.
00:29:44.732 --> 00:29:45.752
not only that, right?
00:29:45.752 --> 00:29:46.792
But the£5 billion of the
00:29:47.767 --> 00:29:48.317
Yes.
00:29:48.563 --> 00:29:50.212
is contingent upon them doing
00:29:50.416 --> 00:30:09.059
Yes, which is also just incredibly strange because there's an element where it's like, okay, if Meta fails and cannot get YouTube and TikTok to agree to this, which also includes getting them to contribute a similar amount of money to the states, then Meta doesn't have to pay the$5 billion.
00:30:09.059 --> 00:30:18.231
So Meta's in this weird position where it's like it can save$5 billion from this agreement if YouTube, and TikTok don't agree to this.
00:30:18.361 --> 00:30:29.883
But, from the competitive side of it, you know that Meta wants to get them to do it because, as you pointed out, kids spend more time on YouTube and TikTok than they do on Instagram.
00:30:29.942 --> 00:30:41.932
And so if Meta can convince YouTube to put a two-hour limit on watch time, that actually benefits them because kids will cut off their YouTube watching at two hours, and what else are they gonna do?
00:30:41.992 --> 00:30:43.932
Ah, I guess we'll scroll over to Instagram.
00:30:44.487 --> 00:30:44.916
Yeah
00:30:45.002 --> 00:30:51.215
And so this is like, it's really, really, really weird in how it's structured.
00:30:51.538 --> 00:30:51.798
Yeah.
00:30:52.228 --> 00:30:52.708
And you know what?
00:30:52.708 --> 00:30:58.935
I, I read this and I thought, these guys are playing checkers while everyone else plays chess.
00:30:59.336 --> 00:31:05.066
Like, this is, this is the kind of move that you have to just, like, hold your hands up and go,"Wow, that is...
00:31:05.526 --> 00:31:08.675
No, no one would've thought that that was even possible," right?
00:31:08.705 --> 00:31:12.844
Or, or like that this was a kind of even something that the attorney generals would be up for.
00:31:13.090 --> 00:31:16.827
and it isn't something that has been the focus of the reporting.
00:31:16.837 --> 00:31:28.539
The most of the reporting done by major outlets has been focused on the framing has been kind of as a success for the states and this is a kind of success for the users because there are these limits or are these features that will be disabled at certain points.
00:31:28.976 --> 00:31:32.914
this element has been much less discussed, and I think that's why we wanted to kind of talk about it.
00:31:33.154 --> 00:31:46.721
So they're playing a very interesting game here, Meta, and they've obviously put a price on, their ability to convince YouTube and TikTok to what they call is, create an industry-wide solution, for child safety.
00:31:46.801 --> 00:31:49.742
it's a bit like kind of, you know, if we're going down, we're
00:31:49.817 --> 00:31:50.297
Yes.
00:31:50.353 --> 00:31:51.452
you guys down with us
00:31:51.646 --> 00:31:52.096
Yeah.
00:31:52.326 --> 00:32:03.257
Or the alternative way of looking at it is it's regulatory capture through judicial settlement, which is, you know, a slightly different form of regulatory capture.
00:32:03.257 --> 00:32:10.855
Regulatory capture is usually because there's like a, you know, a law that comes into place, but you help write the law or you help administer the law in some form or another.
00:32:11.174 --> 00:32:19.325
But in this case, it's through the settlement and then sort of saying like,"We're gonna help try and pressure everybody else into this." Therefore...
00:32:19.467 --> 00:32:28.855
if you take a step back, it's basically Meta getting to write the standards for child safety across the big players of the internet.
00:32:29.286 --> 00:32:34.013
And if you don't believe that Meta has been particularly good...
00:32:34.013 --> 00:32:49.019
Like I-- what amazes me of the people celebrating this are the same people who are like,"Meta is terrible, and we don't trust them." And now you have handed them the pen to write the rules of how internet safety for children works across the internet.
00:32:49.436 --> 00:33:07.520
And even if you say,"Well, it's just limited to these biggest players," they can convince YouTube and TikTok to go along with it, the trickle-down of this is also really problematic because now what is happening is you've blessed these specific things, the two-hour limit, the hidden likes, the things that we don't know whether or not they actually work.
00:33:07.686 --> 00:33:32.310
But if any other platform wants to avoid getting sued in the future, the easiest thing to do is just say,"Well, we're doing the industry-wide standard as defined in this settlement between Meta and all of these states." And so the way to reduce liability if you're a company, a smaller company, an upstart trying to do something different, maybe you have an idea of how to actually protect children more.
00:33:32.769 --> 00:33:34.039
That's now risky.
00:33:34.327 --> 00:33:49.028
the easiest thing, the thing to do to reduce and avoid liability is to just go with the flow, take what Meta has agreed to and say,"We're just gonna do that because that is the industry standard." And that really concerns me, especially when we don't know what actually works right now.
00:33:49.453 --> 00:33:49.784
Yeah.
00:33:50.034 --> 00:33:53.648
I think that's a, a, perfect point to end today's free episode.
00:33:53.648 --> 00:34:00.011
I think I, I can see you're getting on, into a roll, and that's a perfect time, I think, to, flip over to our Patreon extended episode.
00:34:00.374 --> 00:34:08.684
listeners, if you want to hear Mike continue his, uh, his thought process, um, use that Trust Con code while you can.
00:34:09.000 --> 00:34:14.429
sign up to become a supporter for either$10 or$20 a month, and, uh, we'll see you on the other side.
00:34:14.429 --> 00:34:15.000
Thanks for listening.
00:34:16.893 --> 00:34:20.173
That's the end of this week's free edition of Ctrl Alt Speech.
00:34:20.554 --> 00:34:27.434
To hear the extended episode with additional coverage, become a Patreon supporter at patreon.com/ctrlaltspeech.
00:34:27.695 --> 00:34:30.233
That's C-T-R-L Alt Speech.
00:34:30.815 --> 00:34:37.235
If your organization is interested in sponsoring the podcast, contact us via our website at ctrlaltspeech.com.
00:34:37.673 --> 00:34:38.393
Thanks for listening
00:00:00.000 --> 00:00:01.627
Mike, I don't know How often you cook,
00:00:02.213 --> 00:00:02.592
Oh.
00:00:04.003 --> 00:00:08.243
But you will have known by now that following a recipe on your phone is an absolute nightmare.
00:00:08.369 --> 00:00:18.414
I've got so bored of like, constant scroll and the skipping between videos and, like, those kind of SEO bait recipes that have become a bane of all of our lives.
00:00:18.672 --> 00:00:22.853
Well, I found a solution, and it's very, very cool.
00:00:22.853 --> 00:00:31.527
It's this thing called Recipe Tables that rather than showing a recipe as a story, has come up with, with what it calls a, prep table.
00:00:31.647 --> 00:00:42.450
And so recipes are now shown visually as a table with the ingredients down one side and the techniques that you need to use to make your meal on the right-hand side.
00:00:42.450 --> 00:00:45.200
So you can see at a glance what you need to do and when.
00:00:45.210 --> 00:00:46.670
You don't have to go back and
00:00:47.012 --> 00:00:47.893
No scrolling.
00:00:48.119 --> 00:00:49.060
it's no scrolling.
00:00:49.079 --> 00:00:49.939
It's fantastic.
00:00:50.390 --> 00:00:59.582
Um, and this tool allows you to paste any URL into its tool, and it kind of displays your recipe, exactly in this kind of special prep table form.
00:00:59.593 --> 00:01:00.383
So it's wicked.
00:01:00.482 --> 00:01:01.103
It's really cool.
00:01:01.442 --> 00:01:01.923
Very nice.
00:01:01.923 --> 00:01:05.763
It's very old-fashioned though, because the kids these days are using AI to help them.
00:01:06.272 --> 00:01:10.718
Just ask the AI to tell them what other ingredients they need for their recipe.
00:01:10.992 --> 00:01:11.263
True.
00:01:11.263 --> 00:01:11.783
Yeah, true.
00:01:11.783 --> 00:01:13.162
What have I, what steps have I missed?
00:01:13.462 --> 00:01:13.492
Yes
00:01:14.643 --> 00:01:16.069
what have you been looking at this week?
00:01:16.411 --> 00:01:21.355
Uh, so the, the really cool thing that I saw on the internet that reminded me of how cool the internet can be...
00:01:21.408 --> 00:01:23.207
do you know who Rainbolt is?
00:01:23.847 --> 00:01:25.378
There's a YouTuber named Rainbolt.
00:01:25.408 --> 00:01:26.418
Have you ever heard of him?
00:01:27.048 --> 00:01:32.918
So his sort of claim to fame is that he is very, very good at the game GeoGuessr.
00:01:32.957 --> 00:01:34.287
Have you ever played GeoGuessr?
00:01:34.477 --> 00:01:34.867
Yes.
00:01:34.878 --> 00:01:36.275
So very fun, whatever.
00:01:36.275 --> 00:01:43.564
So Rainbolt is sort of a, a YouTube influencer who is very good, though my children tell me he is not the best at GeoGuessr.
00:01:43.564 --> 00:01:47.834
He is merely a good YouTuber at GeoGuessr.
00:01:48.051 --> 00:01:54.840
but occasionally, you know, he'll, he'll do other kinds of videos, but most of them around his sort of skills at finding stuff.
00:01:55.350 --> 00:02:05.775
So he put up this video that I warn you ahead of time may make you cry, because someone had reached out to him.
00:02:05.795 --> 00:02:08.745
'Cause, like, people will ask him,"Can you find where this was?" They'll like...
00:02:09.025 --> 00:02:18.594
I've seen him find, like, photos of, like, great-grandparents on their honeymoon in some random place, and he'll figure out where it was, you know, 120 years ago or whatever.
00:02:18.935 --> 00:02:30.764
But this one was, someone said that when they were a kid, they were, like, seven or eight years old or something, and their father had taken them to a Lakers game, basketball, back to
00:02:30.919 --> 00:02:31.000
Ja.
00:02:31.578 --> 00:02:35.617
and that the father had unexpectedly passed away soon afterwards.
00:02:36.057 --> 00:02:42.777
And they, they had a photo from the game, but the kid did not know which game it was and was trying to find which game it was.
00:02:42.777 --> 00:03:05.842
So he gave him the photo and he said,"I promised I would find it." And then he systematically goes through how he used the internet and, like, Getty Images and, news stories, and all of these other, like, little things, and eventually, I'm not gonna, away exactly how he finds it, but at one point he's like, The real secret to all of this were these two blue pixels." And you're just like, oh my God.
00:03:06.342 --> 00:03:08.293
And he figures out which game it is.
00:03:08.383 --> 00:03:30.062
Then he finds online the video of the game, watches the entire video of the game until he spots them, the kid and his father, in the audience, and it's just like, it's, it's really, very, very moving, both incredible about the internet and what can be found on the internet and how it can do it, and then just a very, very moving, ending for that.
00:03:30.112 --> 00:03:36.543
And if you look also, I will say, read the first comment on that video after you've watched the video as well.
00:03:36.562 --> 00:03:36.872
So
00:03:37.187 --> 00:03:37.687
Okay.
00:03:37.796 --> 00:03:39.616
Little Easter egg that you've left us there.
00:03:39.776 --> 00:03:41.977
I mean, that s- that sounds fantastic.
00:03:42.067 --> 00:03:50.759
It's, it's the perfect tonic it sounds like to this, today's episode, where we're gonna, we're gonna be talking about, nothing so wholesome
00:03:51.138 --> 00:03:51.747
Yes.
00:03:52.800 --> 00:03:53.539
unfortunately.
00:04:02.033 --> 00:04:11.122
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.
00:04:11.436 --> 00:04:25.391
It's August 27th, 2026, and this week we're going all in on the big story of the week, Meta's blockbuster online safety settlement, what it entails, what they've agreed, and for our extended Patreon episode, why it matters and what happens next.
00:04:25.627 --> 00:04:38.291
Ctrl Alt Speech is the podcast that, unlike Meta, settles almost nothing, l- leaves almost everything open, and it's co-hosted by Mike Masnick, founder and editor of Techdirt, and me, Ben Whitelaw, founder and editor of Everything in Moderation.
00:04:38.732 --> 00:04:41.278
Mike, this podcast doesn't cost 17 billion pounds.
00:04:41.447 --> 00:04:41.708
Uh,
00:04:42.057 --> 00:04:42.627
Not yet.
00:04:44.108 --> 00:04:45.088
Give us time, Ben.
00:04:45.088 --> 00:04:45.767
Let us cook.
00:04:47.137 --> 00:04:49.305
it's, it's a slow and, gradual growth.
00:04:49.483 --> 00:04:51.682
but yeah, this, this has been the massive story of the week, right?
00:04:51.682 --> 00:04:52.372
It's, it's...
00:04:52.622 --> 00:05:00.485
I like it when there are stories that, we care about that get covered in the mainstream news, and which kind of people will, almost stop you in the street to talk about, right?
00:05:00.485 --> 00:05:01.675
This, this has been one of those.
00:05:01.896 --> 00:05:02.425
Yeah.
00:05:02.475 --> 00:05:11.466
Though, it, I mean, it's, it's crazy because it came out yesterday and, all week long between last week's episode, which was a very good episode you missed, but was still a very good episode.
00:05:11.651 --> 00:05:19.658
and there were a bunch of stories, and I was filling up our, spreadsheet with all the different stories we wanted to cover, and then yesterday they all got wiped out
00:05:20.490 --> 00:05:20.651
Yeah.
00:05:21.379 --> 00:05:23.569
'cause we've got a story to cover today.
00:05:24.019 --> 00:05:24.428
Yeah.
00:05:24.488 --> 00:05:29.369
It was a giant meteor of a story that landed on the Ctrl Alt-Speech Earth.
00:05:29.536 --> 00:05:35.178
we can't really not talk about anything else, and there's so much to it that w- we, couldn't really, even cover it in just a free episode.
00:05:35.178 --> 00:05:38.899
I think we're gonna have to split it out, really break it down, talk through what it looks like.
00:05:39.206 --> 00:05:40.408
we'll get into it basically.
00:05:40.509 --> 00:05:44.569
before we do, you know, the people who haven't settled are our Patreon supporters, Mike.
00:05:44.829 --> 00:05:49.716
They, they are, investing in, in a way that is almost akin to meta
00:05:49.740 --> 00:05:52.610
can we get attorneys general to force people to pay us?
00:05:52.610 --> 00:05:54.029
Is that, is that where we're getting at?
00:05:54.841 --> 00:05:58.346
I think this is a good, this, settlement suggests we might have a case.
00:05:58.874 --> 00:05:59.454
There we go.
00:05:59.836 --> 00:06:00.276
Yeah.
00:06:00.386 --> 00:06:03.326
Um, but we've got, uh, we've got lots of Patreon supporters.
00:06:03.326 --> 00:06:05.685
We've got our free supporters, who we love.
00:06:05.685 --> 00:06:08.286
They're dipping their toe into the waters of Ctrl Alt-Speech.
00:06:08.565 --> 00:06:11.331
they're not paying anything right now, but they're getting a sense of what we provide.
00:06:11.661 --> 00:06:21.262
We've got our supporter tier that are getting our extended episodes and earlier access for, much less than£18 billion, and that is$10 a month.
00:06:21.721 --> 00:06:27.982
And we've got our founder and insider tiers who are getting all that and the chance to suggest stories for$20 a month.
00:06:28.281 --> 00:06:37.478
And as of today, we are rolling out the chance for those supporters to give us ideas about what special episodes we should be doing for our Patreon supporters.
00:06:37.478 --> 00:06:47.336
So if you are an insider or a founder, supporter and you have something you want us to cover, a particular question you have, we are taking ideas for special episodes of Ctrl Alt-Speech.
00:06:47.646 --> 00:06:51.615
We've already done one, which is our s- our reading list episode, which we put out a couple of months
00:06:51.886 --> 00:06:53.475
Which was really good, by the way.
00:06:53.505 --> 00:07:03.935
If, if you haven't listened to it, it was, you know, we each went through three different books that we think are really incredibly valuable if you're interested in the kinds of topics that we talk about.
00:07:04.233 --> 00:07:06.769
and we got some really good feedback on this as well.
00:07:06.809 --> 00:07:14.718
We may end up doing another episode with more books because as we noted, there are a lot of very interesting books in the space, but that was a really, really fun episode
00:07:15.170 --> 00:07:15.670
Definitely.
00:07:15.880 --> 00:07:21.610
I've been questioning whether I included the right three books in that ep- episode, so we will have to do another one.
00:07:21.697 --> 00:07:26.278
but yeah, our Patreon supporters are a massive part of how we produce and sustain Ctrl Alt Speech.
00:07:26.401 --> 00:07:32.074
we also, have some brilliant sponsors who come on board to sponsor and, and reach our excellent audience.
00:07:32.274 --> 00:07:37.913
So if you're interested in becoming a supporter, go to patreon.com/ctrlaltspeech, more information there.
00:07:38.983 --> 00:07:47.141
And if you want to sponsor us, go to the Ctrl Alt Speech website and you can find out more information and some examples of, sponsorships that we've done in the past.
00:07:47.420 --> 00:07:55.197
as we get to the end of August, Mike, as, as August comes to a close, it's important to mention that our Trust Con code is still valid for
00:07:56.093 --> 00:07:57.504
Our discount code, yes
00:07:57.627 --> 00:08:05.468
Our discount code, If you're interested in, doing more than dipping your toe into the waters of Ctrl Alt Speech and you want 90% off your first month, use Trust Con.
00:08:05.870 --> 00:08:10.774
we're there at the end of August doing our live episode, and this is a special promotion we're running.
00:08:11.033 --> 00:08:15.468
that's still available until the end of this month, so ending on Tuesday.
00:08:15.935 --> 00:08:16.485
Yes
00:08:16.769 --> 00:08:17.230
Perfect.
00:08:18.250 --> 00:08:19.610
Glad, glad we're agreed on that.
00:08:20.588 --> 00:08:20.997
No.
00:08:21.548 --> 00:08:22.218
How dare you?
00:08:22.718 --> 00:08:29.447
Uh, yeah but, uh, I mean, we put our all into this podcast, and, we hope that that comes through in what you hear.
00:08:29.747 --> 00:08:45.914
and if you wanna see it keep going and continue to support the work that we're doing in trying to explain these complex topics and make them understandable and make them real, please look into supporting us on, Patreon, and thank you already to those who have already stepped up and support us there.
00:08:46.403 --> 00:08:52.235
go subscribe, and you can do it at a discounted rate if you use the code TrustCon for the next few days
00:08:52.658 --> 00:08:52.948
Indeed.
00:08:53.239 --> 00:09:00.418
And if, if you, want to become a paid supporter right now, you can obviously help us out by rating and reviewing us wherever you get your podcasts.
00:09:00.428 --> 00:09:01.158
That is free.
00:09:01.339 --> 00:09:04.711
all it needs is your time and your, creativity, let's say
00:09:05.148 --> 00:09:08.357
And also, you can promote the podcast to other people.
00:09:08.418 --> 00:09:18.577
You can tell other people about the podcast or about the stories that you liked or that you, we said something that you found interesting or compelling or that changed your mind on something.
00:09:18.618 --> 00:09:19.687
Tell other people about it.
00:09:19.738 --> 00:09:21.717
Word of mouth is how podcasts grow.
00:09:21.758 --> 00:09:23.827
So we always say rate, review, and subscribe.
00:09:23.827 --> 00:09:30.004
That is the sort of official ways, but I think we should be more aggressive in reminding people to tell other people.
00:09:30.004 --> 00:09:35.162
People you know may be interested in what we're saying, and the way they're gonna find out about it is if you tell them
00:09:35.566 --> 00:09:35.936
Yeah.
00:09:36.115 --> 00:09:43.958
Ping that Discord group or that Slack channel or, you know, those friends that you don't really like but think they should have better views on the internet.
00:09:44.219 --> 00:09:46.408
you're the smart person who knows about the internet.
00:09:46.778 --> 00:09:50.489
Do your thing and, uh, share the good word of Ctrl Alt Speech.
00:09:50.918 --> 00:09:55.042
Um, Mike, we really have a lot to get into in so far as this Meta settlement goes.
00:09:55.042 --> 00:09:58.645
This is, a case that's been rumbling since 2023.
00:09:58.895 --> 00:10:01.812
It might have caught a lot of people off guard if they've not been paying attention.
00:10:01.985 --> 00:10:05.525
and so you've written about it extensively this week on Techdirt.
00:10:05.812 --> 00:10:10.408
I've seen a lot of people sharing your post, as a, counterpoint to a lot of the reporting out there.
00:10:10.708 --> 00:10:19.609
But just before we dive into some of those ideas, talk us through what the actual settlement entails, and some of the kind of things that have been agreed off the back of it.
00:10:20.144 --> 00:10:20.514
Yeah.
00:10:20.514 --> 00:10:40.738
So this, this was a case that was bo- brought by most state attorneys general in the US, not all, most of them, against Meta a few years ago, you know, claiming all of the usual things about child safety and, and, you know, the argument being that Meta does not take child safety seriously enough and something needs to be done.
00:10:41.039 --> 00:10:45.238
And the case has been moving forward in sort of fits and starts, over the last few years.
00:10:45.347 --> 00:10:51.727
And, there's the trial that's going on in California right now that was sort of related to that.
00:10:51.817 --> 00:10:59.195
there's obviously also the New Mexico case that we covered, a few months ago when, when all of that happened and, and the rulings there.
00:10:59.475 --> 00:11:10.072
but what happened yesterday as we're recording was that Meta settled with 52 attorneys general, and basically wiped out most of those cases.
00:11:10.317 --> 00:11:11.725
you know, it was a consolidated case.
00:11:11.745 --> 00:11:32.490
But, wiped out that, made an agreement they were gonna pay a bunch of money, that the amount that they pay depends on a few different factors, which we'll get into in a, a moment, they have promised to put in place a bunch of different changes within the app, which are designated to hopefully protect and help protect children in some form or another.
00:11:32.788 --> 00:11:37.807
and they are also going to try to make this into an industry standard.
00:11:37.817 --> 00:11:54.528
The features and changes that they're making, or at least some of the features and changes that they're making, they are trying to impress upon specifically YouTube and TikTok to make the same changes, and there are some incentive structures involved here to get them to try and make it an industry standard.
00:11:54.567 --> 00:12:07.548
And if other social media companies reach to a certain level, then Meta is also effectively required to try to pressure them into implementing the same features as well.
00:12:07.825 --> 00:12:09.404
so that's the sort of headline.
00:12:09.754 --> 00:12:16.528
You know, the headline number that gets a lot of attention is that they're paying, it depends on which reporting you see.
00:12:16.538 --> 00:12:19.418
It's somewhere between$16 and$18 billion.
00:12:20.488 --> 00:12:22.687
and that money will go to the states.
00:12:22.697 --> 00:12:28.087
In theory, that money is supposed to be used for child safety related things.
00:12:28.501 --> 00:12:31.572
it will be paid out over the course of about 10 years.
00:12:31.822 --> 00:12:39.261
you know, so a lot of people, the initial response is Meta is a very large company that makes a lot of money.
00:12:39.534 --> 00:12:43.725
$18 billion over 10 years is not that much
00:12:44.099 --> 00:12:50.947
and actually to put that into context, revenue last year from Meta's various businesses and apps was 200 billion.
00:12:51.087 --> 00:12:54.134
So we're talking about, less than a 10th of last year's revenue
00:12:54.490 --> 00:12:55.809
Spread out over 10 years
00:12:55.957 --> 00:12:58.227
spread, spread out over 10 years, which we, which we talk about.
00:12:58.600 --> 00:13:04.288
they had a particularly good last quarter, I think something like 24 or 25% up year on year.
00:13:04.298 --> 00:13:08.298
So they made 60 billion just in the last quarter alone.
00:13:08.508 --> 00:13:12.260
So it's essentially one third of last quarter's revenue.
00:13:12.691 --> 00:13:17.563
So we're talking about a month's worth of Meta revenue at the current trajectory.
00:13:17.864 --> 00:13:19.043
That's a really important thing to note.
00:13:19.114 --> 00:13:24.114
So in salary terms, you know, I was thinking about it in terms of like my, my salary take home.
00:13:24.254 --> 00:13:26.624
It's a bit like having to replace your boiler,
00:13:27.974 --> 00:13:28.384
Yes.
00:13:28.604 --> 00:13:28.933
right?
00:13:28.964 --> 00:13:29.833
You know, for, for your
00:13:29.903 --> 00:13:30.293
that's a...
00:13:30.403 --> 00:13:35.947
your average person, you know, again, you know, not, not giving away too much about how much I earn here, I don't think.
00:13:35.957 --> 00:13:39.378
But like boiler, bo- boilers vary in, in, in cost.
00:13:39.638 --> 00:13:43.980
But for the average person, it's probably the cost of, an unexpected boiler replacement
00:13:44.418 --> 00:13:44.818
Right.
00:13:45.019 --> 00:13:58.875
you know, I, I think I, I sort of said, it probably stings, but you, it's the kind of sting that you forget about a day later, it is not nothing, but it is something that, effectively will be forgotten by the company very soon
00:13:59.337 --> 00:13:59.508
Yeah.
00:13:59.597 --> 00:14:02.467
And in terms of the usefulness, I mean, boilers are useful.
00:14:02.628 --> 00:14:06.394
We'll talk about how useful this settlement will be for Meta going forward.
00:14:06.394 --> 00:14:08.533
But I think it's important to kind of contextualize that number.
00:14:08.933 --> 00:14:14.870
Talk about the 10 years, Mike,'cause I had to do a bit of, uh, I'm not familiar with a lot of, like, US legal settlements.
00:14:15.250 --> 00:14:17.860
The 10 years kind of matters for a few reasons, right?
00:14:18.288 --> 00:14:22.768
the reason it's spread out, as far as I'm understanding, is that they don't they can't just pay it and forget about it.
00:14:23.138 --> 00:14:23.847
Is that part of it?
00:14:24.500 --> 00:14:27.399
I mean, there's a bunch of moving parts within the settlement.
00:14:27.462 --> 00:14:31.525
and it's not entirely uncommon that you would have this kind of settlement.
00:14:31.596 --> 00:14:35.596
you know, where, where things are paid out over time, where things need to be checked over time.
00:14:35.596 --> 00:14:42.515
There's like, there's a process in which an independent auditor has to get in-involved to make sure that Meta is taking care of these things.
00:14:42.515 --> 00:14:49.296
There are other structures that are being set up as a part of this as well, which is what some of the, the money will get paid into.
00:14:49.655 --> 00:14:53.169
but was there something specific about the 10-year structure that, struck you as
00:14:53.312 --> 00:14:57.231
Well, just as a kind of fairly arbitrary number, um, right?
00:14:57.261 --> 00:14:59.655
You know, it's kind of, I was thinking like why 10 years?
00:14:59.676 --> 00:15:00.591
Why not, five?
00:15:00.591 --> 00:15:02.528
Or why not kind of fewer than, than that?
00:15:02.528 --> 00:15:05.479
But, is it fairly common that these kind of settlements would be over a
00:15:05.899 --> 00:15:12.374
Yeah, this, doesn't strike me-- I mean, you know, this is obviously unique in certain ways, but it's not uncommon for...
00:15:12.433 --> 00:15:26.086
You know, the thing that I would compare the ten-year structure to is the various FTC, consent decrees that basically all of the social media companies ha-have had whenever they've done something bad on the privacy front.
00:15:26.086 --> 00:15:34.466
At some point, they come to a settlement decree with, with the FTC, and those usually stretch over a period of time, which will be ten to twenty years.
00:15:34.787 --> 00:16:00.769
And part of the thinking there is generally like, you know, we wanna make sure that you are continuing to do the things that you are required to do over this time, so we're going to be checking back in, and we're putting in place some rules on, on how you do it, which, you know, there is related to the payments over ten years, the, features that Meta has agreed to, make standard or default, those are required over a period of five to ten years as well.
00:16:00.769 --> 00:16:11.167
So it's, it's sort of stretching out this, this sense of like, know, Meta is committed to doing this and can't roll it back or change things very quickly, after the settlement has been agreed to.
00:16:11.490 --> 00:16:16.303
I don't know that the, payment necessarily directly matters on that front.
00:16:16.303 --> 00:16:22.509
I do think that, you know, it also in some sense keeps Meta engaged, that they don't forget about this.
00:16:22.539 --> 00:16:24.600
Like, it, it almost acts as a reminder.
00:16:25.024 --> 00:16:36.149
on the flip side too, it's like if you're the states and you're getting this settlement, whereas it may be a sort of a drop in the bucket for Meta, it's still a significant amount of money that is suddenly gonna go into the state coffers.
00:16:36.399 --> 00:16:43.500
And if it was all a one-time payment, you know, there may be some benefits to doing it that way, but it's also the kind of thing that easily could then get squandered.
00:16:43.750 --> 00:16:56.850
Whereas if you know that you have a ten-year period where you're gonna get, uh, hundred million dollars a year or whatever, you can sort of plan for that and put that to good use with a sort of expectation of, what's gonna happen over the next ten years.
00:16:56.947 --> 00:17:01.376
and so I think the hope is that if you do it that way, the money will actually put to good use.
00:17:01.543 --> 00:17:08.304
you know, we'll see how the states eventually actually use it or how much transparency it-- there is into how the states use it right now.
00:17:08.574 --> 00:17:10.854
I don't know that there is that much transparency.
00:17:11.166 --> 00:17:12.876
I mean, I think that would be really interesting.
00:17:12.886 --> 00:17:18.396
Are the states actually going to put this into child safety in any meaningful way, and what does that mean?
00:17:18.406 --> 00:17:19.227
What does that look like?
00:17:19.227 --> 00:17:21.527
Are the states in agreement of how they're gonna use the money?
00:17:21.876 --> 00:17:26.420
That's information that as far as I know, haven't been told yet, and I don't know if we'll ever be told.
00:17:26.794 --> 00:17:26.993
Right.
00:17:27.213 --> 00:17:27.534
Yeah.
00:17:27.564 --> 00:17:28.034
Interesting.
00:17:28.054 --> 00:17:28.344
Okay.
00:17:28.624 --> 00:17:32.557
so that covers off the amount and the kind of time span that this settlement is over.
00:17:32.938 --> 00:17:49.153
Talk us through the kind of what Meta has agreed to, um, as part of this then, because, yeah, I, I read in your tech, uh, post that you felt it, to some degree that what they've agreed to in terms of product changes actually doesn't really matter, but it's still what's being reported fairly heavily.
00:17:49.412 --> 00:17:49.832
Yeah.
00:17:49.862 --> 00:18:00.271
I mean, it matters to users of Meta's b- uh, products, especially the teen users, because some of the changes are things that aren't really changes.
00:18:00.294 --> 00:18:08.526
A lot of them are things that Meta was kind of already doing or already embracing or already pushed into by other legal processes over, over the last few years.
00:18:08.892 --> 00:18:29.419
so there are a bunch of things, and they're all sort of features that, if you read them, you might say, like,"Oh, okay, these are things designed to, protect teenagers." the big one on, my side is the, age assurance, the guarantee of doing age assurance, which age assurance is the polite way of saying age verification, which is the polite way of saying identity verification.
00:18:29.788 --> 00:18:44.885
age assurance, they always sort of hand wave around it where it's like,"Oh, if we're doing age assurance, we don't necessarily need to get your ID." But generally these things lead back to you have to provide some sort of identifi- the service in some form or another.
00:18:45.281 --> 00:18:55.827
I suppose there's a lot of, uh, margin for error in age assurance, which kind of looks at a biometric image or some other kind of signals, right?
00:18:55.877 --> 00:18:56.288
Yeah.
00:18:56.386 --> 00:19:16.028
some of the issue is just, like, if you're having, like, hard cliffs within your age categorization system, if it's like 16 and under has to be treated this way, the general age assurance is not going to be that good in that the age assurance system might be able to tell the difference between a 12-year-old and a 60-year-old.
00:19:16.382 --> 00:19:18.929
Anyone can tell the difference between a 12-year-old and a 60-year-old.
00:19:19.199 --> 00:19:36.067
Where it gets very tricky is when you're trying to tell the difference between, like, a 14-year-old and a 17-year-old, and if that is important to the systems that you're putting in place, and that can matter because the types of things, experiences, and content that are appropriate for a 14-year-old could be very different than for a 17-year-old.
00:19:36.163 --> 00:19:54.663
and so if you're trying to judge down to that very granular level of, like, between 15 and 17, you generally will need some sort of identification to really prove who someone is, and then you're getting into the question of, like, privacy and the rights of teenagers.
00:19:54.693 --> 00:20:01.818
And as we discussed last week when you were away, France just threw out their, social media ban.
00:20:01.847 --> 00:20:11.487
the constitutional council there threw out the ban that, Macron put in place because they said, like, it is ignoring the fact that teenagers have rights, too, and they do.
00:20:11.608 --> 00:20:21.573
you know, and I've said in the past, like, when platforms are deciding to put in place age-appropriate systems on their own, that is perhaps more understandable.
00:20:21.573 --> 00:20:33.019
But when it's, at the behest of the state, then it begins to get tricky and begins to tiptoe into areas where it's like now the government is trampling on the rights specifically of teens.
00:20:33.257 --> 00:20:45.756
The other aspect of it is that when you're in the age verification, identity verification world, it's easy to say,"Well, you know, we have to protect the kids," but the only way that works is if you identify everybody, right?
00:20:45.760 --> 00:21:03.201
If you're gonna have to know who the kids are, who the teens are, you also have to know who the adults are, which means you're basically doing some sort of process to identify everyone, and then that becomes a sort of big privacy concern over how that is done and how you protect that data and how you keep people safe.
00:21:03.579 --> 00:21:03.890
Yeah.
00:21:03.950 --> 00:21:17.882
And it, and it's worth noting that Meta has in recent months talked about how it is expanding its age assurance measures and using AI to, for example, place teens into its kind of what it calls teen-appropriate experiences.
00:21:17.892 --> 00:21:19.686
So it has been doing elements of this.
00:21:19.894 --> 00:21:23.396
this settlement kind of, I guess, makes that a blanket, thing.
00:21:23.631 --> 00:21:30.300
again, we don't think we know exactly what means they're using, but it's an experience that's gonna be coming to more, more users on Meta.
00:21:30.795 --> 00:21:33.255
so that is one aspect, but there are other aspects as well, right?
00:21:33.255 --> 00:21:40.883
There are lots of kind of little elements that have been given to people who, who are concerned about children's safety
00:21:41.186 --> 00:21:44.386
a whole bunch of other little tweaks and changes.
00:21:44.707 --> 00:21:49.017
The one that I think has probably gotten the biggest headlines is the two-hour time limit.
00:21:49.446 --> 00:21:51.257
So this is across meta platforms.
00:21:51.257 --> 00:21:57.369
Teenagers will be limited to two hours of time per day, and then they'll be cut off.
00:21:57.394 --> 00:21:58.960
there are also regular prompts.
00:21:58.960 --> 00:22:05.173
If you're, like, looking at content for a while, it will remind you you've been looking at this for this much time, sort of leading up to the two hours.
00:22:05.532 --> 00:22:08.252
Now, there are a couple important things on that two hours.
00:22:08.474 --> 00:22:13.587
one of which, which I think is a really big deal, which is that parents can override that.
00:22:14.028 --> 00:22:20.477
and I think there are going to be many stories coming out of, you know, kids are going to be using it.
00:22:20.477 --> 00:22:21.416
They're gonna hit the two hours.
00:22:21.416 --> 00:22:24.836
They're gonna run to the parent and say, like,"Hey, you know, I was in the middle of something.
00:22:24.836 --> 00:22:32.255
I need to get a pass this." And a bunch of parents are going to end up removing that, and then it's not gonna matter, right?
00:22:32.255 --> 00:22:33.845
And so, like, you have all sorts of things.
00:22:33.845 --> 00:22:44.565
There's always issues of, like, when you need parental consent, that raises a whole other, series of concerns about relationships between children and their parents and who has authority and, and whether or not...
00:22:44.964 --> 00:22:55.583
You know, it's great if kids have a good relationship with their parents, and I think it's important for, you know, kids to have parents that they can, communicate with, but not everybody is, so fortunate.
00:22:55.803 --> 00:23:01.911
and you could see where that could lead to conflict and real problems and, issues.
00:23:02.188 --> 00:23:16.932
you know, it has these other features like, it will remove showing you likes because there's this feeling that, like, people start doing things just for the likes, which, is anecdotally very compelling, but I'm not sure that the research actually supports that.
00:23:17.281 --> 00:23:26.190
you know, there's things about, being able to have a non-algorithmic feed, which is something that a lot of people claim is really important.
00:23:26.200 --> 00:23:29.069
But again, the research doesn't necessarily support that.
00:23:29.069 --> 00:23:39.567
In fact, the research has shown that if you, force people into a non-algorithmic feed, they tend to see worse stuff because part of the algorithm is often, finding the better content.
00:23:39.846 --> 00:23:44.896
I will note, going back to my intro, that Rainbolt video was recommended to me by the YouTube algorithm.
00:23:44.926 --> 00:23:45.757
That's how I found it.
00:23:45.876 --> 00:23:48.646
So sometimes the algorithm recommends good stuff.
00:23:49.067 --> 00:23:54.813
Um, so there's, there's a bunch of those kinds of features that, like, people have raised concerns about.
00:23:55.047 --> 00:24:05.192
Where another area, another feature that I think or feature th-thing that I think is concerning is they have, like, uh, specific categories of content.
00:24:05.192 --> 00:24:10.442
So the specific one is disabling cosmetic surgery and extreme makeup filters.
00:24:10.878 --> 00:24:48.182
and they define it fairly narrowly and carefully, and they say, like,"Oh, it doesn't include these other things that, aren't problematic." And, you know, you can look at this and say, Okay, I can understand where there are concerns about, encouraging teenagers to, do drastic cosmetic surgery or things like that." But we're getting into spaces where this is constitutionally protected speech, and as soon as you're in this space where, like, the government has pressured a company into saying, like,"We have to block this specifically defined content from children," it becomes really, really problematic.
00:24:48.553 --> 00:25:00.348
And the issue that I raised in my piece on it is, a lot of these things, like, if you just look at them, y- you say, like,"Sure, like, maybe kids shouldn't have access to this." And I, and I understand why Meta's doing them.
00:25:00.598 --> 00:25:05.740
But many of the feature changes, we don't know if they actually protect children.
00:25:06.144 --> 00:25:13.635
and you can make cases where they probably protect some children, but you can also look at this, this stuff, and I, I've been having discussions...
00:25:13.885 --> 00:25:54.313
Actually, just this morning before we were even recording, I was having a discussion with someone who, who has some expertise on, suicide prevention, and was looking at some of the feature changes and saying, This might be really damaging to kids going through some mental health issues, where if these are not implemented well, it could cause, teenagers dealing with a mental health episode to spiral in some ways." and so there is this element of, people talk about, like, oh, it's unfair or it's bad that companies are experimenting on our children with their algorithmic feeds or the different features and, you know, we don't know if it's safe or whatever.
00:25:54.593 --> 00:26:11.055
But this is also an experiment, and this is because in some back room somewhere, a bunch of Meta lawyers and a bunch of state attorneys general got together and just decided in that room what they think will make kids safer.
00:26:11.681 --> 00:26:23.030
And I don't think that there were probably that many child safety experts in that room, and we don't actually know if any of these feature changes will actually protect kids.
00:26:23.050 --> 00:26:25.270
Will, in some cases, will it make them worse?
00:26:25.760 --> 00:26:27.711
And there doesn't seem to be...
00:26:27.740 --> 00:26:35.768
You know, the settlement agreement does allow Meta to study some of these issues, but, you know, they're still sort of locked into them.
00:26:35.998 --> 00:26:46.425
And there is, like, an independent auditor that is a part of the process, but the independent auditor is to, audit that Meta is actually doing these things, not that they're working or protecting children in any way.
00:26:46.607 --> 00:26:46.968
Yeah.
00:26:46.968 --> 00:27:06.036
Well, there is this, independent social media research foundation, which it sounds nice, but there is not a lot of detail about what information would be available to researchers, academics, whether that would be historical data, live data, and whether it would kind of chronicle, this period under which these changes will be made.
00:27:06.256 --> 00:27:07.801
So I think that, that's a fair call out.
00:27:07.801 --> 00:27:14.641
I, I also wanted to kind of point out, I guess, on the time limits, two hours a day across Meta products sounds good.
00:27:14.921 --> 00:27:21.570
I think we, we had this idea that, teenagers are using their phones a lot on social media, that that maybe is a reduction.
00:27:21.830 --> 00:27:32.084
I was looking up, one poll, one Gallup poll in advance of chatting, and, and there's apparently boys and, girls between the age of 13 and 18 spend about five hours on social media.
00:27:32.395 --> 00:27:35.997
And Meta are not the top most used
00:27:36.230 --> 00:27:39.611
I was gonna say it's not Meta that they're spending that time on.
00:27:39.717 --> 00:27:40.186
right?
00:27:40.317 --> 00:27:47.027
So, so according to this one poll based upon 2,000 US teens, YouTube is, is around two hours.
00:27:47.836 --> 00:27:51.999
TikTok is, around, an hour and a half across boys and, and girls.
00:27:52.138 --> 00:27:53.269
Instagram is an hour.
00:27:53.298 --> 00:27:55.481
Facebook is, 3 of a, of an hour.
00:27:55.501 --> 00:28:08.830
So and it will come onto this, and I think it's relevant because we- we're gonna talk about Meta becoming an industry standard and how they are, they've made some contingencies within the settlement, that relate to their competitors, YouTube and TikTok.
00:28:08.830 --> 00:28:11.451
So I think this two hours sounds like a, a lot.
00:28:11.809 --> 00:28:14.558
it maybe isn't as, as little as we thought
00:28:15.295 --> 00:28:15.595
Yeah.
00:28:15.625 --> 00:28:30.863
I mean, when you look at it, and then we'll get into what the requirements are for other platforms, but yeah, somebody pointed out, like, this would not actually probably diminish that much usage of Meta products because most people don't use them more than two hours.
00:28:30.863 --> 00:28:33.212
It is the combined with other platforms.
00:28:33.722 --> 00:28:42.730
But, you know, if you look at the other platforms, the YouTube and the TikTok, a lot of that interaction and engagement is a different type, right?
00:28:42.881 --> 00:28:46.451
YouTube is more sort of sit back and watch engagement.
00:28:46.461 --> 00:28:55.048
That can be true of, Instagram, but Instagram is l- usually a little bit more engaged and involved and, reading, replying, all that kind of stuff.
00:28:55.348 --> 00:28:57.288
TikTok, I think, is somewhere in between.
00:28:57.338 --> 00:29:02.419
You know, it is a lot of just sort of watching videos, but also some level of engagement.
00:29:02.852 --> 00:29:06.211
and I think YouTube is-- tends to be more just sit back and watch.
00:29:06.221 --> 00:29:12.508
More people use YouTube as a kind of modern TV rather than social media in that sense.
00:29:12.728 --> 00:29:21.266
And so, this gets at the thing that, like, all these different platforms are somewhat different, and treating them all as the same can be problematic.
00:29:21.556 --> 00:29:40.434
Which gets to the next concern of this agreement and the weirdest part of the agreement, which is that Meta has to try to convince specifically YouTube and TikTok to put in place the same features, or not all of the same features, but most of the same features, and limits.
00:29:40.684 --> 00:29:44.515
And that is weird on almost every level.
00:29:44.732 --> 00:29:45.752
not only that, right?
00:29:45.752 --> 00:29:46.792
But the£5 billion of the
00:29:47.767 --> 00:29:48.317
Yes.
00:29:48.563 --> 00:29:50.212
is contingent upon them doing
00:29:50.416 --> 00:30:09.059
Yes, which is also just incredibly strange because there's an element where it's like, okay, if Meta fails and cannot get YouTube and TikTok to agree to this, which also includes getting them to contribute a similar amount of money to the states, then Meta doesn't have to pay the$5 billion.
00:30:09.059 --> 00:30:18.231
So Meta's in this weird position where it's like it can save$5 billion from this agreement if YouTube, and TikTok don't agree to this.
00:30:18.361 --> 00:30:29.883
But, from the competitive side of it, you know that Meta wants to get them to do it because, as you pointed out, kids spend more time on YouTube and TikTok than they do on Instagram.
00:30:29.942 --> 00:30:41.932
And so if Meta can convince YouTube to put a two-hour limit on watch time, that actually benefits them because kids will cut off their YouTube watching at two hours, and what else are they gonna do?
00:30:41.992 --> 00:30:43.932
Ah, I guess we'll scroll over to Instagram.
00:30:44.487 --> 00:30:44.916
Yeah
00:30:45.002 --> 00:30:51.215
And so this is like, it's really, really, really weird in how it's structured.
00:30:51.538 --> 00:30:51.798
Yeah.
00:30:52.228 --> 00:30:52.708
And you know what?
00:30:52.708 --> 00:30:58.935
I, I read this and I thought, these guys are playing checkers while everyone else plays chess.
00:30:59.336 --> 00:31:05.066
Like, this is, this is the kind of move that you have to just, like, hold your hands up and go,"Wow, that is...
00:31:05.526 --> 00:31:08.675
No, no one would've thought that that was even possible," right?
00:31:08.705 --> 00:31:12.844
Or, or like that this was a kind of even something that the attorney generals would be up for.
00:31:13.090 --> 00:31:16.827
and it isn't something that has been the focus of the reporting.
00:31:16.837 --> 00:31:28.539
The most of the reporting done by major outlets has been focused on the framing has been kind of as a success for the states and this is a kind of success for the users because there are these limits or are these features that will be disabled at certain points.
00:31:28.976 --> 00:31:32.914
this element has been much less discussed, and I think that's why we wanted to kind of talk about it.
00:31:33.154 --> 00:31:46.721
So they're playing a very interesting game here, Meta, and they've obviously put a price on, their ability to convince YouTube and TikTok to what they call is, create an industry-wide solution, for child safety.
00:31:46.801 --> 00:31:49.742
it's a bit like kind of, you know, if we're going down, we're
00:31:49.817 --> 00:31:50.297
Yes.
00:31:50.353 --> 00:31:51.452
you guys down with us
00:31:51.646 --> 00:31:52.096
Yeah.
00:31:52.326 --> 00:32:03.257
Or the alternative way of looking at it is it's regulatory capture through judicial settlement, which is, you know, a slightly different form of regulatory capture.
00:32:03.257 --> 00:32:10.855
Regulatory capture is usually because there's like a, you know, a law that comes into place, but you help write the law or you help administer the law in some form or another.
00:32:11.174 --> 00:32:19.325
But in this case, it's through the settlement and then sort of saying like,"We're gonna help try and pressure everybody else into this." Therefore...
00:32:19.467 --> 00:32:28.855
if you take a step back, it's basically Meta getting to write the standards for child safety across the big players of the internet.
00:32:29.286 --> 00:32:34.013
And if you don't believe that Meta has been particularly good...
00:32:34.013 --> 00:32:49.019
Like I-- what amazes me of the people celebrating this are the same people who are like,"Meta is terrible, and we don't trust them." And now you have handed them the pen to write the rules of how internet safety for children works across the internet.
00:32:49.436 --> 00:33:07.520
And even if you say,"Well, it's just limited to these biggest players," they can convince YouTube and TikTok to go along with it, the trickle-down of this is also really problematic because now what is happening is you've blessed these specific things, the two-hour limit, the hidden likes, the things that we don't know whether or not they actually work.
00:33:07.686 --> 00:33:32.310
But if any other platform wants to avoid getting sued in the future, the easiest thing to do is just say,"Well, we're doing the industry-wide standard as defined in this settlement between Meta and all of these states." And so the way to reduce liability if you're a company, a smaller company, an upstart trying to do something different, maybe you have an idea of how to actually protect children more.
00:33:32.769 --> 00:33:34.039
That's now risky.
00:33:34.327 --> 00:33:49.028
the easiest thing, the thing to do to reduce and avoid liability is to just go with the flow, take what Meta has agreed to and say,"We're just gonna do that because that is the industry standard." And that really concerns me, especially when we don't know what actually works right now.
00:33:49.453 --> 00:33:49.784
Yeah.
00:33:50.034 --> 00:33:53.648
I think that's a, a, perfect point to end today's free episode.
00:33:53.648 --> 00:34:00.011
I think I, I can see you're getting on, into a roll, and that's a perfect time, I think, to, flip over to our Patreon extended episode.
00:34:00.374 --> 00:34:08.684
listeners, if you want to hear Mike continue his, uh, his thought process, um, use that Trust Con code while you can.
00:34:09.000 --> 00:34:14.429
sign up to become a supporter for either$10 or$20 a month, and, uh, we'll see you on the other side.
00:34:14.429 --> 00:34:15.000
Thanks for listening.
00:34:16.893 --> 00:34:20.173
That's the end of this week's free edition of Ctrl Alt Speech.
00:34:20.554 --> 00:34:27.434
To hear the extended episode with additional coverage, become a Patreon supporter at patreon.com/ctrlaltspeech.
00:34:27.695 --> 00:34:30.233
That's C-T-R-L Alt Speech.
00:34:30.815 --> 00:34:37.235
If your organization is interested in sponsoring the podcast, contact us via our website at ctrlaltspeech.com.
00:34:37.673 --> 00:34:38.393
Thanks for listening
