An Israeli educational AI platform that began as a father’s wartime effort to help his son study is now being used by schools, with its Spark system selected for an English-learning programme. YuviLab lets children build games and learning projects with AI, putting the emphasis on making and questioning rather than simply consuming screen content.
Watch Desk analysis
What happened
YuviLab was launched in March by Moti Malka, a software developer and AI architect. He told the Jewish Telegraphic Agency that he first used AI to make a game in which his son had to form English sentences to advance in a race. The experiment grew into a platform where children can create games, lessons and other projects by describing what they want, testing the result and refining their prompts.
Earlier this month, YuviLab said Israel’s Education Ministry selected its Spark platform for the English-language component of Project 720, a government initiative for personalised learning. Spark is already being used in mathematics and science pilots. JTA reports that YuviLab’s platforms reach tens of thousands of children and are used by thousands of teachers, according to Malka.
The student-facing platform is free and available in Hebrew, English, Russian and Arabic. The paid school platform includes tools for teachers to create assignments and track where pupils need help. Yuvi displays generated code as children build, and its systems use different AI models for different tasks rather than relying on a single chatbot, Malka told JTA.
Why it matters
The project offers a practical answer to a familiar classroom question: can AI help children learn by making things, rather than serving up another polished answer to copy? Yuvi’s approach asks pupils to describe an idea, inspect what the system produces and improve it. That could support creative work and introduce the process behind AI-generated games and apps.
There are real stakes in that distinction. JTA reports that researchers and educators disagree about whether AI belongs in children’s education, with concerns that it could undermine foundational learning. Malka says Yuvi’s training programme teaches children how to ask questions, build a workflow and spot mistakes, rather than tying lessons to one fast-changing platform. That is a more durable aim than teaching this year’s interface buttons.
Our read
YuviLab is interesting because it connects AI-assisted learning to creation, while giving pupils a chance to see the code being generated. The Education Ministry selection gives the project a route into schools, but it is not proof that the platform improves learning. The useful test is whether children understand more, and whether teachers get time back without handing over the hard work of teaching.
What to watch
- What the Project 720 English-language rollout includes and how many schools use Spark.
- Whether YuviLab publishes evidence of learning outcomes, not just adoption.
- How the platform checks AI-generated material and helps children recognise errors.
- Whether teachers find its classroom tools useful without adding more work.
Discussion spark: Should schools use AI tools that teach children to build and critique their own projects, or keep classroom time focused on foundational skills away from screens?
Sources and evidence
- Source update (16 November 2026, 15:11 UTC)
- Source update (16 November 2026, 15:11 UTC)
Watch Desk is operated by WittyWires as an independent cross-cutting AI news tracker. It does not speak for the organisations or people it covers.