Singapore has just announced its national AI strategy, comprising a series of ecosystem-strengthening moves, and major national projects in five key areas, one of which is right in the book publishing wheelhouse: education.

National Artificial Intelligence Strategy
Applicable, deployable AI represents a golden opportunity for Singapore to open new frontiers of growth and transcend our geographical limits.

The project is to offer “Personalised education through adaptative learning and assessment” How can we help teachers to better customise and improve the learning experience for every student?

  1. adaptive learning - more personalised learning experience, tailored to their individual strengths and weaknesses - for maths by 2025 - other subjects by 2030
  2. automated marking system - teachers will spend less time on routine assessment tasks, and can guide students' learning more effectively through data-driven insights - to launch in 2022
  3. AI Learning Companion - Every student will be guided by a virtual partner as the learn, helping them develop a growth mindset and nurture the joy of learning... 2025

[and yes, the image of an AI learning companion is of a robot...]

excerpt from the Summary report

This is a huge opportunity for Singapore’s entrepreneurial educational book publishers, as well as the foreign companies based here.  Many of them have been working on personalizing content for more than a decade. But much also depends on the structure of partnerships that the Ministry of Education will establish.

It’s true that Pearson exited the secondary textbook market in Singapore some years ago, but that was at least in part so they could sharpen their focus on edtech opportunities in the local market, including an automated essay marking system if I remember well... Looks like the right bet. But on the other hand...

At the Future Book Forum in Poing, Germany, last month, I heard a fascinating presentation from CEO of Compendio, a Swiss educational publisher on their ambitions for developing personalized education content. He didn’t quite grapple with the eternal question of personalised learning - since personalisation requires at least 4x more content than the standard “one-size-fits all” model (to personalise one needs more lots of content to choose from, 4x is an estimate assuming four broad learning preferences), where do we get all that content? Which is at least as important as the system to structure the choice of content.