Towards Autonomous Learning: a thought experiment inspired by cars

Buckle up, kids.

Will Jackson
3 min read5 days ago


Autonomous learning, students at the wheel (generated with Midjourney)

Do you love driving — the feeling of control, the thrill of the open road? Or do you dread rush hour traffic and dream of the day your car can drive itself all the way to your destination? Perhaps a little bit of both — so you may have mixed emotions about autonomous vehicles becoming a reality in the next 10–15 years.

To be specific, Level 4 autonomy (the vehicle can handle most road situations without human input) is expected to become commonplace in the 2030s, according to McKinsey. The Levels of Driving Automation classify vehicle autonomy from Level 0 (no automation) to Level 5 (fully autonomous driving). And while Level 4 autonomy can already be seen on the streets of San Francisco, many experts now doubt that Level 5 autonomy will ever be technically or commercially viable.

Despite the differences between driving and teaching, I wondered whether a similar ‘levels’ framework would be helpful to frame discussions about adopting AI within K12 education. I found the thought experiment to be both inspiring and a little scary — join me and decide for yourself on a journey into the Levels of Learning Automation.

At Level 0: No Automation, teachers are fully in control. They create lessons, instruct, set homework, grade work, and provide feedback. The students are passengers in this teacher-led journey. Student motivation comes from teacher recognition and extrinsic reward — gold stars and grades — supplemented by peer interaction and intrinsic reward.

Teacher, take the wheel (generated with Midjourney)

Transition to Level 1: Driver Assistance, and AI comes into play, assisting teachers with tasks like lesson planning and preparation. Students start interacting with AI during structured activities, but motivation remains largely driven by teachers, peers and intrinsic interest.

By Level 2: Partial Automation, AI starts to suggest personalized learning objectives and resources. It helps with assessments and facilitates peer collaborations. Teachers, while still at the helm, now approve AI-made choices. Students begin to follow an AI-tailored learning path, motivated by a growing sense of personalized learning and achievement.

At Level 3: Conditional Automation, based on its detailed understanding of aptitude and psychology, the AI curates and creates personalized materials for students, suggests learning paths, convenes and moderates student learning communities and grades digital assignments. Teachers help students to select their learning paths, provide emotional support and step in for complex inquiries.

With Level 4: High Automation, AI manages the learning process, facilitates virtual projects, and provides real-time feedback. Teachers are now learning supervisors, stepping in to provide coaching in response to social and emotional cues. Students feel in control of their learning navigation, but the AI is nudging students towards a learning path based on aptitude and assessed opportunities, on behalf of the student’s school or LEA.

Finally, we would reach Level 5: Full Automation. Here, AI would evolve student learning experiences based on real-world trends and the student’s life changes. Teachers are matched by AI with students based on pedagogical and psychological need, and provide human connection and coaching in response to social and emotional cues. Students are motivated by tailored goals, peer communities, and real-world project connections. The AI is working to optimize societal objectives of full employment, national security and social harmony.

How does this thought experiment make you feel — Sceptical? Cynical? Scared? Excited? All of the above? Me too! This thought experiment represents not just the adoption of technology in education, but a reimagining of roles and responsibilities, and raises questions about technology’s power over human lives and human autonomy.

As with Vehicle Autonomy, it’s by no means assured that Level 5 is an economically or socially viable outcome. But…my bet is that something like Level 5 will be the outcome in some societies, while others will settle into Level 3 or Level 4. What do you think?

This article was written with help from Chat GPT Plus (Level 1).



Will Jackson

Will is CEO and co-founder of mote, and lives in the San Francisco Bay Area.