Chess has enjoyed something of a moment in the spotlight in recent months, as blockbuster show The Queen’s Gambit captivates audiences around the world.
This modern media renaissance is just the latest chapter in a history of a game that stretches back over 1,500 years, with the widely accepted origins of chess attributing its emergence to the 6th Century CE. This beloved game has enjoyed a remarkable history since, lauded as a defining test of intellect in many countries around the world.
The link between chess and intelligence goes even deeper than one might think however. For centuries chess has been at the heart of humanity’s attempt to unlock automation. The reality is that chess has played a foundational role in our development and understanding of automation, artificial intelligence (AI), and machine learning (ML).
First moves in automation
The first move in chess automation was a remarkable opening gambit. In the late 18th century, German inventor Wolfgang von Kempelen constructed his seemingly-wondrous automated chess playing machine — the Mechanical Turk — to wow audiences in Austria.
This complex machine hinted at unlocking a new era of automation, as it successfully competed against human chess players. It was a bluff. The contraption was an elaborate hoax, allowing a human player to hide inside the machine and operate controls that moved the pieces. Despite its audacious origins, this machine was an early foundation of automation’s love story with chess.
In 1914 the first truly automated chess machine was unveiled at the University of Paris. El Ajedrecista (The Chess Player) was designed by Spanish engineer Leonardo Torres y Quevedo. It was an automaton capable of playing the endgame of a chess match. As well as being a revolution in automation, it is widely viewed as the first computer game in history.
El Ajedrecista operated on a very simple algorithm, which while rather rudimentary in its chess playing, was revolutionary for its time. It not only undertook the correct moves to win the endgame at every playing, it had also been programmed to recognize and flag illegal moves by opponents.
Entering an era of modern chess intelligence
Chess automation entered the contemporary computing landscape in 1950, led by none other than the father of modern computing himself, Alan Turing.
Turing created a code for an automated chess program in 1950, six years before the term ‘artificial intelligence’ was first brought into existence at a conference at a Dartmouth College workshop. He named his algorithm ‘Turbochamp’, and while its complexity meant it never ran on computers of the time, it was finally given the chance to shine in June 2012, when the University of Manchester gave it life as part of its Alan Turing Centenary Conference.
By this stage chess had firmly been established as a popular measure of AI ‘intelligence’. That position would continue over the coming decades, as institutions and researchers across the world sustained their fascinating investigations into AI.
In 1957, the first complete chess program was created on an IBM. It reportedly took eight minutes to make a move. Various iterations emerged over the next decade. In the 1960s MacHack was created in a lab at MIT. It became the first chess program to beat a player at a tournament, and earn its own rating.
By the 1980s chess computers were entering a new era of power. Belle, devised by Bell Labs, became the first machine to achieve master status. It dominated the world of computer chess. Then something more powerful came along to rock the chess world, lifting AI to a new level of success.
A new master of AI
In the early ‘90s, a new computing chess champion emerged in the form of IBM’s Deep Blue. This powerful chess AI utilized parallel computing technologies, creating a supercomputer of unrivalled power.
The series of challenges between Deep Blue and reigning world champion Gary Kasparov are now legendary. The first such contest in 1996 ended in defeat for the computer. Kasparov reigned supreme.
In 1997, a rematch was scheduled. The latest iteration of Deep Blue ‘met’ Kasparov in the Equitable Center in New York, where, unfazed by the gaze of the world media, the chess computer beat the reigning human champion over a six-game series. This not only marked a seminal moment for computing intelligence, but a valuable public recognition of the remarkable advances in modern computing technology.
The truth is that the most advanced human master in chess history would have no chance of beating today’s leading computing brains. The architecture of precursors such as Deep Blue is now fundamental to research in everything from drug development to financial modeling. These advances paved the way for complex AI and machine learning solutions like Neural Technologies’ own Optimus platform, which leverages advanced data modelling techniques to deliver solutions in revenue protection, digital transformation, and customer engagement.
The evolution of automation and AI engagement with chess continues to today, despite its now unquestioned dominance over human players. In 2017, Google repurposed an AI designed for the game of Go into a chess-playing computer. The new system — AlphaZero — was introduced to chess with no prior understanding of the rules. Within four hours it had mastered the game and beaten world champion chess program Stockfish 8 over 100 games. This fundamental machine learning approach acquired understanding with self-reinforced knowledge.
The history of chess and chess AI offers a remarkable look at the evolution of computing technologies. While this relationship is one to stir public attention, the true value of this evolution is almost unseen in our expectation of computing technologies today. In the question of chess, AI now reigns supreme. While man may have lost the battle on the board, the remarkable discoveries progressed by this research remain a fundamental driver of our modern connected world.