Demis Hassabis proposes moving rote learning to AI and turning classrooms into project hubs
The Google DeepMind chief executive argues schools should focus on teamwork and critical thinking while personalised AI handles basic memorisation at home.
Schools need to stop spending class time on memorisation and rebuild their days around project work, group collaboration, and human judgement, according to Google DeepMind chief executive Demis Hassabis.
Speaking on how education must adapt as artificial general intelligence develops over the next five to ten years, Hassabis argued that the traditional one-size-fits-all classroom model is already obsolete. Instead of teachers lecturing to an average speed that leaves quicker students bored and struggling students behind, AI tools should deliver custom lessons at home.
What Demis Hassabis proposed for schools
Under the proposed model, memorising formulas, vocabulary, and core facts would move out of the school day entirely. Students would complete that work at their own pace using software tailored to their exact learning speed.
Time inside the school building would instead focus on group dynamics, entrepreneurialism, and creative problem-solving. Hassabis pointed to Montessori-style group tasks and Oxbridge-style tutorial sessions as the templates mainstream schools ought to adopt at scale.
"School should be about people skills, projects, entrepreneurialism, creativity, judgment, all the things that you might expect the skills to be important in five, ten years' time when AGI's arrived," Hassabis said.
Why traditional lecturing needs to change
Beyond shifting memorisation to home study, Hassabis questioned why millions of teachers around the world spend time creating identical lesson slides for standard courses such as secondary school mathematics. He described the current repetition across individual school departments as completely redundant.
He suggested that a select group of world-class educators — "almost like superstars" — should record and broadcast primary lectures for widespread use. Local teachers would then be freed from prep work to concentrate on mentoring, managing group projects, and supporting individual student needs.
| Learning element | Traditional model | Hassabis proposal |
|---|---|---|
| Rote learning | Classroom drills for all | Personalised AI software at home |
| Class speed | Set to the group average | Adapted to each individual |
| Core lecturing | Prepared by every teacher | Broadcast by top specialists |
| School hours | Direct instruction and tests | Group projects and mentorship |
Differing views on technology in class
The vision outlined by Hassabis sits within a wider debate over how much access children should have to screen-based tools. While some tech executives, including Duolingo chief executive Luis von Ahn, have similarly argued that automated tutors will eventually outpace human teachers at individualized instruction, others favour strict limits.
British historian Niall Ferguson has argued for restricting access to electronic devices during school hours, contending that spending time away from screens is essential to safeguard independent thought.
Hassabis, however, maintains that locking AI out of education misses the point. By delegating routine memorisation to software, he argues, schools can finally turn class hours into spaces dedicated entirely to human interaction.
Key numbers
| What | Figure | Source |
|---|---|---|
| Estimated AGI arrival timeline | 5 to 10 years | Demis Hassabis |
In this story
| Name | Role in the story |
|---|---|
| Demis Hassabis | Chief Executive Officer of Google DeepMind who proposed the classroom reform. |
| Google DeepMind | Artificial intelligence research laboratory headed by Demis Hassabis. |



