Hello!

Hope you're enjoying the excitement that this variable weather is bringing us.

In honour of the mixed-up weather, we're sticking to a mixed-up format. One top story this week, three "also in the news" items, and a new section I'm calling "What happened to...?" where we check back in on some of the stories we've covered in the past. And, as ever, we finish up with our blog spotlight.

So, read on for takes on the under-16s social media ban, a police officer caught fabricating evidence using AI, a banned AI model, a new phone from an old company, and a revisit of electronic shelf labels and coffee-based controversies.

And for our blog spotlight this week, we look at eBooks vs paper books.

Let's dig in...

Top Story

⛔ UK Government announces social media ban for under-16s

You can read the government fact sheet here: Fact sheet: New rules to protect children online

A quick summary

Our main story this week is the UK government's decision to ban social media for under 16s. this has been under consultant for a while, and this week Starmer announcing that from "spring 2027" sites including Snapchat, TikTok and Facebook will be inaccessible for under 16s. Additionally, there is talk of introducing an "overnight curfew" for those aged between 16-18, though at the time of writing it doesn't appear as though this has been given a whole lot of additional thought.

How will this work?

We aren't sure yet, although it seems likely that the same option taken by other countries will apply. Those people with accounts online where the setup occurred over 16 years ago will be exempt, but for the rest some form of government-ID linked verification will likely have to take place.

We've already covered why this is a bad idea (see the story on the Reddit fine in LT3 #19), but more recently our friends over at the EU aren't having the best time with their shiny new government backed, super secure answer to the ID verification problem. By which I mean it was hacked almost instantly. If you count "changing a very easily accessible plain text file that was storing sensitive data on the device" to be "hacking".

Whatever tool the government opts for, it is likely to come hand in hand with be some form of restriction placed on the accessibility of VPN services - one of the easiest and most common ways to circumvent a particular country's rules.

Will it make children safer?

In short: "no".

There's a number of angles to this, but we can start with the facts we already know from the guys down under. 60% of children in a country with a social media ban are still accessing social media (LT3 #31). Starmer is undeterred by this, and says that our system will be "Australia +". Though he, so far, hasn't been overly forthcoming with details on what that means, other than to say the measures will be "highly effective", and ordering Ofcom to conduct an urgent review of age-verification tech.

But lets say he's right. Let's say we develop a system that isn't fooled by pencilled on moustaches, that is immune to editing of text files, that allows people with a genuine safety need for anonymity online to go about their business, and that actually stops all under 16s from accessing the likes of Facebook, TikTok, and Snapchat. What then?

Among the many, many people and organisations who are concerned by this outcome is Dr Catherine Knibbs (PhD), who points out that while the government is treating social media like an illicit substance, in actuality it's more like an environment. It's a place where people exist, not simply a product they consume. Conversation, learning, social connection, all happen in these online spaces. In a recent TikTok video, she argued that we don't ban children from shopping centres just because you might find child predators there. Instead we put the onus on the shopping centre, on our policing system, and on basically everyone except the children, to make sure they can't be harmed.

She's not alone in this assessment. the Office for the Children's Commissioner for Jersey also points out that social media companies "should be ensuring that their platforms are safe by design and creating appropriate spaces for children's engagement". Child Rights International Network - among many, many others - are worried about the "cliff edge" that 16 years olds would face being thrown into a social media landscape that hasn't been forced to reckon with inherent toxicity, rather just temporarily ban its most vulnerable.

Even if it all goes right, there's another worry

But even if all of these people are wrong, and everything Starmer promises comes to pass, I have yet another concern.

I'm not being dramatic when I say that, as a 30-something year old, this ban terrifies me. It's not just the "what", but the "how". And I don't mean the identity capturing debate. I mean the legal mechanism by which this ban will be implemented. Something that gets missed in a lot of the tech coverage of the Online Safety Act (OSA) is that now it's been passed, the government can start to put in place laws without all the bother of Acts of Parliament. In some cases - such as this one - they can rollout regulations and legislation practically unfettered.

Section 214A of OSA gives the Secretary of State for Science, Innovation, and Technology the ability to make new regulations "for the purpose of protecting 'relevant children' from a risk of harm". As a result, rather than new laws around child safety online requiring an Act of Parliament (and the associated debates, amendments, tabled motions, scrutiny from both Houses, weighing-in from experts, etc), the government gets to send the law to a basic yes/no vote. Combine that with a three line whip and Labour's 402/650 majority, and Liz Kendall and Kier Starmer can do pretty much whatever they like.

So we shouldn't bother?

To be clear, it doesn't seem as though anyone thinks a more considered approach to social media for children is a bad idea. In fact, WMG Academy in Coventry is reporting that kids are becoming much more social as a result of smartphone bans. It chimes with a story we ran way back in LT3 #10, where over half of children asked said they'd prefer to live in a world without social media. We need a healthier approach to online spaces. Even The Bear's Jeremy Allen White thinks so...

But trying to ban across a whole generation is a different ask to banning within the physical and temporal confines of a school day. It cuts across too many things, invalidates too many people's privacy, and fundamentally ignores the actual problem with unsafe online spaces - the goddamn unsafe online space.

What else is happening in tech?

👮‍♂️ Derbyshire officer investigated over alleged use of AI to fabricate evidence in multiple cases

There are plenty of stories these days about people going to court with AI-fabricated evidence to try and win their cases. Usually it's the members of the public. Recently in Derbyshire though, the police also wanted to get in on the fun.

An unnamed police officer has been removed from front-line duties as a result of an accusation of deliberately fabricating evidence using AI tools. Naturally, details are thin on the ground, and Derbyshire Police are not likely to be more forthcoming at least until the end of the investigation. However, it seems as though AI was used to create evidence that was later relied on in court for "several" open cases.

While this is the first reported case of its kind in the UK, it seems unlikely that this hasn't happened before. In a telling part of an interview with the Financial Times, head of the National Police Chiefs' Council's PoliceAI Alex Murray, said that some police forces had been told to stop using AI to help turn interviews into court statements. Regardless of the apparent deliberate nature of this particlar case, if AI has been used as widely as is reported then it's almost impossible to imagine that some level of AI hallucination hasn't made its way into the criminal justice system.

🤖 Anthropic withdraws access to powerful AI model after US government order

Hey, remember when Anthropic announced Mythos, and it was so powerful that Mozilla had to fix 271 bugs practically overnight, the NHS closed up all its software, and 4,000 GitHub repositories were compromised? (LT3 #28, LT3 #30) Well they've released it to the public. Sort of. Claude Fable 5 is the latest public release from Anthropic, and uses Mythos as it's base model - though the company is quick to point out that Fable has been heavily guardrailed so as not to pose a cyber security risk.

The US government didn't agree, and within hours of Fable 5's release had demanded that Anthropic suspend access. While Trump's administration specified that only "foreign nationals" should be banned (including, by the way, US-based Anthropic employees who are foreign nationals), Anthropic made the decision to pull access to Fable across the board.

No word yet on when it will be reinstated. Apparently the government wasn't specific about the nature of the qualms, which I imagine will make adjusting the guardrails more challenging.

📞 Commodore announce new flip phone

Relax, you haven't fallen through a time vortex into a parallel world. Yes there is a news story referencing both flip phones and the manufacturer of the legendary 1982 Commodore 64 - to this day the best selling computer system of all time.

While the OG Commodore company shut up shop in 1994, the name has been bought and sold a couple of times (once with an interesting amount of deception), most recently by Retro Recipes YouTube channel owner Christian Simpson. Their retro-inspired "Callback" clamshell phone is an almost fully featured smartphone. The bits that are missing? The social media, and a browser. You can still get WhatsApp, Uber, Maps, email clients and in fact "99% of Android apps", you just can't get your social media fix.

Commodore aren't messing about with this either. Not only can you not download social media apps from the "Commostore" (perfect app store name, no notes), you also can't sideload apps onto the device. And for the super-nerds who find a way around that, at a phone hardware level they've blocked access to the likes of the TikTok servers. (Perhaps Mr Starmer might consider buying all UK kids one? With somewhere around 2.5 million 13-15 year olds in the UK, and a handset cost of £380, the less-than-£1-billion price tag makes it only slightly more expensive than filling up the car.)

The phone's on pre-order at the moment, but I'll be willing to bet that once it kicks off properly, we won't see the last of this kind of thing - a middle ground between a full "dumbphone", and the all singing and dancing AI native flagships. As people start to lose more and more trust in big tech, and the horrors of social media become more and more apparent, maybe people will lean into a phone that just does... what you need a phone to do? Take some decent photos, pop them in a WhatsApp group, and listen to a podcast on the way to work.

Oh, speaking of which, it has a headphone jack. Helpful as wired headphones are also making a comeback.

What happened to...?

Welcome to a new section, where every [amount of time redacted] we'll revisit a story from a previous newsletter to see what the latest is.

Two for you this week:

🛒 Tesco rolling out electronic shelf labels to 3,000 stores

First seen in: LT3 #28 ("Bank of England suggests electronic shelf labels might lead to dynamic pricing in supermarkets")

What's new?
When we last checked in with the story, Morrisons, Co-op, Waitrose and Asda were all rolling out ESL tech in their stores. Now, Tesco have joined in with 3,000 of their "Extra", "Express", and "Metro" stores, in the largest ESL project in Europe.

Tesco PR are pushing the "paper saving" angle, alongside staff efficiencies (not having to spend time manually changing price labels). Interestingly, they have also said that the ESLs will allow other features, such as displays that can light up to help staff pick the correct items for online orders. In order to do that, presumably the ESL will need to know who is in front of it, and what they are shopping for - both prerequisites for some of the personal pricing things we spoke about a few weeks ago.

☕ Starbucks Korea to temporarily shut all stores for history lesson after bungled coffee promotion

First seen in: LT3 #33 ("How a Starbucks marketing stunt spiralled into mass boycotts in South Korea")

What's new?
This one didn't talk long to come back round at all... Just last week we spoke about Starbucks Korea's AI-planned marketing disaster. Now, the new CEO has approved the shutdown of almost all of its 2,000+ stores next week for "social sensitivity training", a series of recorded lectures on contemporary Korean history. The training is expected to cover how companies should account for historical and social sensitivities in their marketing.

It seems unlikely that the average barista will have much say on national marketing campaigns, but the company is hoping that this drastic demonstration of contrition will win back customer loyalty.

The move is expected to cost Starbucks Korea the equivalent of £1m, but comparatively a small price to pay against fallen sales, which still sit at around 25% below pre-"Tank day" levels, over a week later.

Blog spotlight

📖 Physical vs. eBooks

A nice short one from bibliophile Amanda this week, on the merits of ebooks vs paper books. With Amazon discontinuing support for a vast number of older Kindles a little while back, I'm seeing more and more people jump onto the discourse of how best to do one's reading.

Amanda isn't talking about the Kindle debate specifically (though does reference Kobo readers and DRM), but her whole point does speak to an ongoing shift away from essentially renting access to our media (Netflix for our films, Spotify for our music, etc etc), and back to a more traditional form of ownership.

For me, it was yet another good reminder that - while I love shiny new tech - we don't need it everywhere.

And we're out! How are you getting on with the format changes? Do you have a preference? Let me know!

In the meantime, thanks as ever for reading, and I'll see you next week.

Will