Every child will get an AI tutor, but not a human teacher

The author's argument is grounded in a personal story: about 15 years ago, a young man who looked homeless came to the author's IT school, offered to do any job to raise the money, completed the course, taught there for a while, and got his life back on track. The author says he made it that far thanks to teachers who believed in him and motivated him not to drop out — something an AI tutor could not supply.

The piece also points to limits of AI-only learning. The author notes that Sal Khan, who promised a personal AI tutor for every student on the planet, admitted this year that most students weren't using it. And it cites a Wharton School experiment in which about a thousand Turkish high school students practiced math with regular ChatGPT, a tutoring version, or no AI: during practice both bots improved results, but on the exam without bots the regular-ChatGPT group fell behind the no-AI group by 17%, having gotten used to hints.

The context is a teacher shortage and budget pressure. UNESCO predicts the world will need an additional 44 million teachers by 2030, with 4.8 million in Europe and North America, and the U.S. alone has 56,000 unfilled teaching positions. Teachers are the largest expense for American public schools — nearly 30% of all costs go toward teacher salaries — and many districts are looking for cheaper alternatives. The author warns that AI's rapid spread, combined with this shortage, could lead society to accept a reduced teacher role, even as the U.S. Department of Education this year made AI development a grant priority over commentators' objections that it would reduce teacher interaction.