The concept of artificial intelligence (AI) can technically be traced back as far as the 1300s, although the idea wouldn’t be brought to life for centuries still. On this week’s special extended edition of Throwback Thursday, we’re looking at the history of artificial intelligence and what took it from a theoretical framework to something fully embedded into modern society. Intrigued? Wanting to learn more? Let’s get started…

Although the concept of machine learning can go back further, our journey today starts in Madison Square Garden in 1898, with Nikola Tesla (whom we talked more about here). He demonstrated the world’s first radio-controlled vessel. Interestingly, the boat was equipped with what Tesla described as “a borrowed mind”. A somewhat eerie and other-worldly term that came about before the word “robot” was coined in 1921 by Czech writer Karel Čapek. The word was introduced in his play R.U.R. (Rossum’s Universal Robots), where “robot” derives from the word “robota”, meaning work.

Sticking to the world of science fiction, we look to the 1927 film Metropolis. The picture features a robot double of a peasant girl named Maria, who unleashes havoc upon Berlin in 2026. This was the very first robot to be depicted on film, and even inspired the look of Star Wars’ C-3PO.

Japan has built a reputation for unparalleled robotic engineering, so it probably doesn’t come as a surprise that the first robot was built there. In 1929, Makoto Nishimura designed Gakutensoku, which in Japanese means “learning from the laws of nature”. It had the ability to change its facial expression and move its head, using an air pressure mechanism.

Between 1943 and 1950, some very influential and important papers were published. In 1943, Warren S. McCulloch and Walter Pitts published “A Logical Calculus of the Ideas Immanent in Nervous Activity”. This influential paper discussed networks of artificial neurons and how they could perform simple logical functions. This went on to inspire neural networks and Deep Learning. The year 1949 saw two publications worthy of mention -- Edmund Berkeley’s Giant Brains: Or Machines That Think and Donald Hebb’s Organisation of Behaviour: A Neuropsychological Theory. Both touched on the link between the humanistic brain and machine learning or “thought”. The next year, 1950, was a big year for theoretical development, thanks to Claude Shannon’s Programming a Computer for Playing Chess, the first published article on developing a chess-playing programme, and Alan Turing’s Computing Machinery and Intelligence. This is where the infamous “imitation game” was proposed, later becoming the “Turing Test”.

In 1952, Arthur Samuel put theory into practice and developed the first programme to play chess, thus the first computer programme to learn by itself. This was yet to be coined with the term “artificial intelligence”, which came three years later in 1955, from the proposal for a “two-month, 10-man study of artificial intelligence”. The workshop that followed the proposal in 1956, led by John McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon, is popularly considered to be the official birth of the new AI field.

Going into the 60s, AI started to flourish. Computers had the ability to store more, became faster, cheaper and more accessible to those who didn’t have $200,000 a month to lease one (the cost in the early 1950s). The first industrial robot, Unimate, was put to work in 1961 on an assembly line in a New Jersey plant. In 1965, Joseph Weizenbaum developed ELIZA, which carried on a dialogue in English on any topic. To the surprise of Weizenbaum, many of the people who interacted with ELIZA attributed human-like feelings towards the programme.

In 1970, the first anthropomorphic robot, WABOT-1, was built at Waseda University in Japan. The robot had a limb-control system, a vision system and a conversation system. However, despite so much technological advancement, further developments in AI revealed mountains of obstacles. The main issue was the lack of computer power and storage; computers essentially lacked the ability to store the right amount of information and process it fast enough. Patience and funding both dwindled, leaving research to come almost to a standstill for the next decade.

The 80s saw the revival of AI interest from two powers -- a boost in funds and an expansion of the algorithmic toolkit. WABOT-2 was created, and this time the humanoid robot had musical abilities. It could communicate with a person, read a musical score and play tunes on an electric organ. The Japanese Ministry of International Trade and Industry contributed around $400 million to the Fifth Generation Computer Project (FGCP), hoping to develop computers that could converse, translate languages, interpret pictures and generate reason. The decade also saw the first driverless car -- a Mercedes-Benz van! Unfortunately, the goals of the FGCP weren’t met and funding ceased, leaving AI to once again slip away from the spotlight.

However, in the absence of government funding and general public hype around it, AI still managed to thrive. Over the next decade or so, from the 90s onwards, a lot of goals were achieved that simply weren’t possible until this point. Richard Wallace developed chatbot A.L.I.C.E. (Artificial Linguistic Internet Computer Entity), inspired by the ELIZA programme. In 1997, the world chess champion of the time and grandmaster, Gary Kasparov, was defeated by Deep Blue, the chess-playing computer programme. The same year saw the development of speech recognition software by Dragon Systems, which was implemented on Windows. The next year, 1998, saw the first domestic pet robot, Furby! Not only was the Furby a 90s/00s icon, but also the first fully accessible robot that found its way into millions of homes worldwide.

If we skip forward a little and look back over the last decade, we have advancements in AI that couldn’t have been dreamt of 100 years ago. In 2014, Google’s self-driving car passed a U.S state self-driving test in Nevada. In an age of “big data”, AI has stepped in where the human brain fails to process huge sums of information. It really comes into its own in industries, such as technology, banking, marketing and entertainment. We’re on track to achieve artificial intelligence that surpasses human capability completely. Think sentient robots from movies, although we’re hoping it doesn’t go the route of I, Robot, the Gunslinger from Westworld or the aforementioned Maria from Metropolis.

We hope you’ve enjoyed this extended cut of Throwback Thursday. Let us know what you think by getting in touch via social media, or if you have any questions regarding Goodfellow’s product line, contact a member of the team today.