Personally, I wouldn't trust anything generated by AI, particularly for something like this where definitions have changed and are changing, and the amount of rubbish about diabetes on the internet appears infinite.
Bilous and Donnelly in the Handbook of Diabetes (5th ed, published in 2021) say
A steady increase (2.5–3% per annum) in the incidence of type 1 diabetes has been reported worldwide, especially among young children <4 years old. There are large differences between countries in the incidence of type 1 diabetes, e.g up to 10‐fold differences among European countries.
I'd strongly suspect that a ten-fold difference between countries might have a lot more to do with failure to identify and diagnose than an actual incidence difference. If you're not actively looking for something it's probable you won't find it - so a system assumption that T1 is mostly diagnosed in children will be confirmed by the system only looking for T1 in children.