HomeArticlesMachine Learning & Neural Networks

Donald Michie: From Codebreaking to Machine Learning

Donald Michie's pioneering work in machine learning, starting with the ingenious MENACE device, laid the foundations for modern AI techniques and highlighted the importance of understandable artificial intelligence.

mysimulator teamUpdated June 2026≈ 3 min read▶ Open the simulation

In 1960, Michie created MENACE (Matchbox Educable Noughts and Crosses Engine)

In 1960, Michie created MENACE (Matchbox Educable Noughts and Crosses Engine). It consisted of 304 matchboxes, each representing a possible state of a Tic-Tac-Toe board.

Inside were colored beads representing moves. By removing beads that led to a loss and adding beads for a win, the machine 'learned' to play perfectly over time.

This was a tangible demonstration of Reinforcement Learning, decades before AlphaGo.

This was a tangible demonstration of Reinforcement Learning, decades before AlphaGo. Michie went on to found the Department of Machine Intelligence and Perception at the University of Edinburgh, which became one of the world's leading AI centers.

live demo · related simulation● LIVE

He was also a pioneer in robotics, building 'FREDDY', a robot that cou

He was also a pioneer in robotics, building 'FREDDY', a robot that could assemble wooden blocks and cars using vision and touch. Michie championed the idea that AI should be understandable to humans, a concept now referred to as 'Explainable AI' (XAI).

Frequently asked questions

What did Donald Michie’s playful yet profound experiments demonstrate about machine learning?

Donald Michie’s playful yet profound experiments demonstrated that the mechanism of learning could be simple, even if the result was intelligence.

Who created and maintains the UK AI Knowledge Base?

The UK AI Knowledge Base is maintained by an unnamed organization. High-Quality Educational Content.

Try it live

Everything above runs in your browser — open Decision Tree Live and change the parameters while it is running. Nothing is installed, nothing is uploaded, the whole model lives in one tab.

▶ Open Decision Tree Live simulation

What did you find?

Add reproduction steps (optional)