How well can AI map education debates across countries?
Education debates generate more material than anyone can easily follow. We tested what AI could help us see: what is discussed in each country, how issues are framed, and how the debates differ and change over time.
We collected public documents from education media, government, parliament, unions, civil society and research organisations. We ran a separate topic model for each country, then named and grouped the topics into shared categories for comparison.
This section shows what AI brings into view, how far it reflects what we already know and what it misreads or misses.
Where does this evidence come from?
See who published the documents we collected and how the mix differs across the three countries.
What is discussed, and how does it change?
Explore the topics identified in each country, how much of each collection they account for and how they change over time.
Can topics be compared across countries?
See how 138 country-specific topics were grouped into 20 shared categories, and where apparently similar issues differ.
Can AI help us keep up with the debate?
Inspect monthly LLM snapshots and explore the trade-off between speed and completeness.