Artificial Intelligence

The conversation about artificial intelligence in education alternates between two equally useless registers: the enthusiasm that promises universal personalisation, and the alarm that predicts the replacement of teachers. Neither helps someone who has to decide what to do next year with a real budget.

I write from a specific position: that of someone who spent years trying to make information systems work inside government, and who knows how much the implementation — not the technology — determines the outcome. AI applied to bad data produces bad decisions faster. That is the concrete risk, and it gets less attention than it deserves.

Recurring themes: what the evidence currently supports about AI and learning, model governance in the public sector, AI in educational data analysis, and the questions an administrator should ask before signing a contract.