#3 - Fable 5, sovereign AI, and a smarter way to teach kids data privacy
If you go to any gathering right now and bring up AI, people usually talk about the same two things - forgetting a more important one.
You can’t escape the AI conversation. No matter where you go these days, if AI comes up - whether it's at a tech meetup or a kids's playground - I promise you will hear one of the two things:
- It’s going to steal our jobs.
- Look at this beautiful poem it wrote for my dog.
But there is a much bigger shift happening behind the scenes that has nothing to do with how "smart" AI is.
I didn't even get a chance to try the Claude Fable 5 when the US Government made Anthropic suspend it.

What we need to think more about is the control problem: who has the off-switch? (aka sovereignty)
History shows us that relying entirely on someone else for a vital resource is dangerous. When you don't control your own supply, your future can be held hostage during a crisis.
Look at the 1973 oil crisis. When OPEC cut off oil exports over a geopolitical dispute, global economies were instantly crippled by skyrocketing prices and gas shortages. The lesson was brutal but clear: never let another country hold the plug to your economy.
That single event triggered a massive race for independence. The US began aggressively investing in its own domestic energy, and France pivoted so hard toward nuclear power that it still runs on 70% nuclear energy today.
These days, data and computing power have become just as vital as oil. Relying on foreign tech monopolies introduces massive risks regarding data privacy, sudden service cutoffs, or extraterritorial laws (like the US CLOUD Act, which allows US law enforcement to demand data stored by US companies anywhere in the world).
Now, we also have AI - and the entire world is sleepwalking into that exact same dependency trap with it. Everyone seems surprised that some European countries and Australia, for example, are building their own LLMs (Large Language Models).
Australia's aim is simple: Sovereign AI. They are actively building domestic models trained on local data and hosted entirely on their own soil. Why? Because they realize that if an entire country relies on a single tech giant in Silicon Valley for its everyday automated banking, legal research, and government infrastructure, they don't actually control their own future.

If you are still treating Large Language Models (LLMs) like you used to treat Microsoft Word in the 90s - as a simple productivity tool - you are doing it wrong. When you integrate an AI agent into your core supply chain, your customer service, or your security systems (including national security!), it stops being a utility. It becomes critical infrastructure.
If the entity that owns that infrastructure decides to change their terms of service, raise prices by 10x, or, due to geopolitical tensions, geofence your entire country out of their API, your operations don't just "slow down." They stop.
Trustworthy, responsible AI is resilient. The Fable 5 shutdown proved that tech infrastructure is no longer insulated from geopolitical crossfire: a single government order or policy shift can instantly cut access to the brain of your application. If you don’t own the infrastructure of your intelligence, you don’t get to decide when the lights go out.
Comics over lectures: a smarter way to teach kids data privacy
It can be tough to get kids to engage with serious topics like online safety, privacy, and digital literacy - textbooks and lectures often go in one ear and out the other. That is exactly why educational webtoons - a type of digital comics that originated in South Korea and are usually meant to be read on smartphones - are a great solution.
To be honest, I did not know what webtoon means. I had to look it up! And I think Luxembourg's National Commission for Data Protection (CNDP) did really with it's recently unveiled Leonora platform.
Just look at the topic the webtoons explore!


There is also a game kids can play with their mates as a team (with joining facilitated through a QR code). Looks fun - and it does inform teens about many aspects of data privacy pretty well.
It targets kids aged 12-16, but I suspect you could try it with younger children too.
And, like pretty much everything that comes from Luxembourg, it's available in English, French, German and Luxembourgish. Check it out and share your (and your kids') thoughts!
Off-duty
This is the part where we shift gears and look at life outside the day job: it's usually as a mix of screen-free hobbies, cool sci-fi movies, fascinating reads and podcasts (some about AI, because let’s face it, it’s just part of our day-to-day life now), and general decompression.
The definite highlight of the past week was Cirque de Soleil I saw for the first time in my life. For me, it was a celebration of human body and its possibilities, but also - the music, the artistry of it, our whole family was amazed. If you ever get a chance to go, absolutely do not miss it - it is pure magic.

What I am reading
This is not entirely off-duty, though - but it concerns everyone, regardless of whether they work in tech or not.
"The energy required to generate a typical AI image is enough to power a 10-watt LED bulb for 17 minutes, and the energy required for a high-complexity AI video is sufficient to run that same bulb for 42 hours. Similarly, the electricity-associated water footprint is about two tablespoons (29 mL) for a single image, but jumps to 4.1 liters for a complex video—almost equivalent to a two-day drinking water need for one person."

What I am listening
This podcast - Daily AI Brief - has been my running companion for a long time, and I have just found out that the website is brilliant, too. Here is the page for this episode, with great visual representation of key ideas and quotes.



It provides a full transcript in .md format too - "feed it to your agent". Love it.
What I am watching
If you need a good stretch after a workday - or before! - I have recently discovered a very good yoga channel - instead of a yoga studio, the sessions are held in parks, with beautiful scenery as a backdrop.
Thanks for being here and reading along! Your feedback means a lot to me. What did you think of today's issue? Just hit reply to share your thoughts - I read and reply to everyone.
- Let's connect: if we aren't LinkedIn official yet, let's change that!
- If you liked this issue, please forward it to a friend! It’s the best way to help this newsletter grow.
- If you were forwarded this email, you can subscribe to the newsletter here.
Happy Monday - and have a great week!
Katya