Top 10 Biggest AI Breakthroughs in History

The idea of AI didn’t suddenly appear with the advancement of technology. In fact, the concept of automation and machine learning is the result of centuries of deep, philosophical debate involving some of philosophy’s biggest players. It was first coined in the early 17th Century by famous French philosopher René Descartes in his book Discourse on the Method. 

In the book, Descartes makes the first distinction between what we now know as general AI software capable of adapting to any job, and specialised AI programmed for a specific task. He questioned that: “even though some machines might do some things as well as we do them, or – they would inevitably fail in others.” His book also asked questions about AI that still boggle the minds of industry experts today: Can a machine be human? Can it think like a human?

At the start of the 20th Century, the idea of AI had captured the minds of science fiction writers, filmmakers and scientists around the world. In 1927, the sci-fi film Metropolis was released featuring the first AI robot. In 1950, a collection of predictive short stories by Isacc Asimov hit the shelves, creating an imaginary world where robots control society. 

Artistic imagination and excitement for the future brought AI and robotics to the attention of huge names in the scientific community. In 1950, science pioneer Alan Turning described what began known as “The Turning Test” – a test for measuring when we can finally declare that machines can be intelligent. His test was straightforward: if a judge cannot differentiate between human and machine, could the machine trick the judge into thinking they are human? As one of his bold predictions about the future of computing, turning believed that the machine would pass his test by the end of the twentieth century. It seems his prediction was some twenty years. slightly premature. Generative AI chatbot ChatGPT, launched in November last year, is reportedly the first AI system to pass the test.

In 1949, Donald Hebb wrote The Organisation of Behaviour, an academic work pointing out that neural pathways become stronger each time they a used, a fundamental concept to understanding how humans grow and learn. As computing advanced in the 1950s, this realisation gave birth to the idea of simulating a hypothetical neural network using computers. 

The first AI-wired neural network was created in 1951. It was dubbed “SNARC” (the Stochastic Neural Analog Reinforcement Computer), created by Marvin Minsky and Dean Edmonds using vacuum tubes, motor engines and clutches. The machine was tasked with the simple challenge of helping a virtual rat navigate a maze. The machine would send instructions to navigate the maze while documenting its actions using its vacuum tubes. This allowed the machine to learn and shift the probability of success autonomously, giving it a greater chance of findings its way out of the maze. In essence, it is essentially the same process used by Google’s AI systems to identify objects in photos today.

Before Alexa, Siri, and ChatGPT, there was ELIZA – the world’s first chatbot wired by AI. Created by Joseph Weizenbaum from 1964 to 1966 at the MIT Artificial Intelligence Library, ELIZA was designed to imitate a Rogerian therapist by rephrasing many of the patient's statements as questions and asking them to the patient. The chatbot worked best if users limited their conversation to their emotions, but was one of the first AI technologies to truly capture a human’s voice in writing. 

Eliza influenced a number of early computer games by allowing for additional kinds of interface designs. Don Daglow designed an enhanced version of Eliza on a DEC PDP-10 minicomputer at Pomona College in 1973 before writing the computer role-playing game Dungeon in 1975.

By the early 1970s, the excitement surrounding AI was fading as the millions of dollars invested into development failed to produce results that met investors’ high expectations. But this all changed in the 1980s when the XCON expert learning system from Equipment Corporation was credited with saving the company $40 million annually from 1980 to 1986. 

This was a major breakthrough for AI development because it showed that AI software was not just an example of technological innovation, but also a business tool with real-world applications. It paved the way for AI’s employment in the enterprise as large corporations began to comprehend the potential impact AI could have on business operations and how investment into AI software could help save money in the long run. By 1985, investments in AI programmes had risen to $1 billion.

Though AI-wired decision-making with neural networks had existed for some time, it wasn’t until the late 1980s that AI researchers shifted their approach to AI advancement from a rules-based approach to one based on statistics. Experts say the shift can be traced back to IBM’s paper “A statistical approach to language translation”, which championed statistical-based decision-making for the automated translation between French and English languages. 

IMB fed into their system 2.2 million pairs of sentences in French and English to train its translation system using transcripts taken from transcripts from the Canadian parliament. The process would eventually be imitated by Google for Google Translate, except Google has the entire internet at its disposal rather than just a series of transcripts.

One of the major events to popularise AI was when world chess champion Garry Kasparov was publicly defeated by IBM’s Deep Blue supercomputer in what has become one of the most famous chess matches of all time. The win allowed the global population and not just those involved in the AI industry to understand the rapid advancement and growing capabilities of computing and AI software. 

Deep Blue was able to win by processing thousands and possible moves every second based on the data of thousands of earlier professional chess games. “He’s playing the ghosts of the grandmaster's past,” IBM explained in response to the shocking win.

After a period of radio silence from artificial intelligence, while the world embraced smartphones, tablets, and accessible internet at home in the early 2010s, AI took the world by storm again with the launch of Apple’s AI-wired Voice Assistant, Siri. Siri’s voice recognition engine uses advanced machine learning algorithms to function, with its original American, British and Australian voice actors recording their respective voices in 2005. 

Many praised Siri’s ability to process and decipher casual language and deliver specific and accurate results from questions and commands, which is something AI assistants had previously struggled with. Today, Siri is one of the most high-profile applications of machine learning. Along with Google’s Assistant, Microsoft’s Cortana, and Amazon’s Alexa, Siri changed the way people interact with their devices forever.

Autonomous vehicles have been in development since the 1950s, but it wasn’t until 2018 that the self-driving car officially hit the road with the launch of Waymo’s self-driving taxi service in Phoenix, Arizona. The taxis were driven by “Waymo Driver,” an AI-wired driving software that uses data from sensors to decipher its surroundings while creating highly detailed, custom maps of over 20 million miles in simulation. The software then uses an AI to analyse the data it gathers from its maps to create the safest and safest route to its destination. 

What made Waymo’s taxis so revolutionary is that, unlike other autonomous vehicles before them, their employment was not limited to testing and experimentation. As many as 400 individuals paid to be driven by Waymo’s driverless cars when the technology first launched, demonstrating the commercial demand for the technology.

When OpenAI’s ChatGPT launched in November 2020, its capabilities shocked the world. The generative AI chatbot collects huge amounts of data from the internet to create accurate, human-like responses to almost any query in the world and almost any language. It can write, essays, articles and poems, translate text and generate ideas from nothing – from creating a medical research plan to planning a child’s birthday party. 

The AI technology is so sophisticated that it was able to pass an MBA exam, so accurate that online news outlets used it to write articles, and so innovative that it has grabbed the eyes of tech behemoths like Microsoft – who has already invested billions. Experts within the field have suggested that generative AI technologies like ChatGPT could mark the beginning of a transformation of society that implements AI technologies like never before. It remains unclear where ChatGPT will take the world of AI next, but the future looks promising.