Artificial intelligence

 WHAT IS ARTIFICIAL INTELLIGENCE?

Artificial Intelligence, commonly abbreviated to AI, is the simulation of human intelligence processes by computer systems. These processes include learning, reasoning, and self-correction. The term was coined by American computer scientist John McCarthy in 1956.

The word "artificial" here means made by humans rather than occurring naturally, and "intelligence" refers to the ability to acquire and apply knowledge. Whether machines truly "think" in any meaningful sense remains one of the most contested questions in philosophy.

Image by Perplexity

EARLY HISTORY

The idea of artificial beings endowed with intelligence goes back to ancient times. Ancient Greek myths described Talos, a giant bronze automaton built by the god Hephaestus to protect the island of Crete.

Alan Turing is widely regarded as the father of theoretical computer science and artificial intelligence. In 1950, Turing published a paper called Computing Machinery and Intelligence, which opened with the question: "Can machines think?" He proposed what became known as the Turing Test — if a human cannot tell whether they are conversing with a machine or a person, the machine can be considered intelligent.

John McCarthy, who coined the term "Artificial Intelligence", was also the inventor of the programming language LISP, which became the dominant language for AI research for decades. He initially called his 1956 conference proposal the "Dartmouth Summer Research Project on Artificial Intelligence" — a title even he later admitted was longer than it needed to be.

The first AI program is generally considered to be the Logic Theorist, created in 1955 by Allen Newell and Herbert Simon. It was designed to mimic human problem-solving skills and was able to prove 38 of the first 52 theorems in Whitehead and Russell's Principia Mathematica.

AI WINTERS

Despite early enthusiasm, AI research ran into two prolonged periods of reduced funding and interest, known as "AI winters." The first lasted through much of the 1970s, the second through the late 1980s and 1990s.

During the first AI winter, the U.S. and British governments drastically cut funding to AI research after a report by mathematician Sir James Lighthill concluded that the field had failed to live up to its promises. Researchers were not pleased.

FAMOUS AI MILESTONES

In 1997, IBM's Deep Blue chess computer defeated world chess champion Garry Kasparov in a six-game match. Kasparov accused IBM of cheating, suggesting a human had helped the computer make one particularly inspired move. IBM denied this and dismantled Deep Blue shortly after, which Kasparov found suspicious.

In 2011, IBM's Watson appeared on the quiz show Jeopardy! and defeated two of the show's greatest champions, Ken Jennings and Brad Rutter. Jennings wrote beneath his final answer: "I for one welcome our new computer overlords."

Google DeepMind's AlphaGo became the first computer program to defeat a professional human player at the ancient board game of Go in 2015. Go had long been considered a game too complex and intuitive for computers to master. AlphaGo's victory came a decade earlier than most experts had predicted.

In 2022, a chatbot called ChatGPT, developed by the company OpenAI, reached one million users in just five days of launch — faster than any other consumer product in history. For comparison, it took Netflix three and a half years to reach the same milestone.

AI IN EVERYDAY LIFE

The autocorrect function on smartphones is a form of AI. It is responsible for many of the world's most embarrassing text messages.

Spam filters in email accounts use AI to decide what is junk mail. Early spam filters were fooled by spammers who inserted the word "not" before every offensive word. Modern filters are harder to trick.

The algorithms that decide which films Netflix recommends to you, which posts appear on your social media feed, and which advertisements follow you around the internet are all forms of AI. Netflix has estimated that its recommendation algorithm saves the company over $1 billion per year by keeping subscribers from cancelling.

ODD FACTS

The term "robot" predates the concept of AI by several decades. It was coined by Czech writer Karel Čapek in his 1920 play R.U.R. (Rossum's Universal Robots). Čapek himself said he would have preferred the word "labori" but was talked out of it by his brother.

The first AI to pass the bar examination — the American lawyers' qualifying test — was GPT-4, in 2023. It scored in approximately the 90th percentile. Many humans do not pass the bar at all.

ELIZA, one of the earliest chatbot programs, was created at MIT in the mid-1960s. Its creator, Joseph Weizenbaum, was disturbed to find that people — including his own secretary — were forming emotional attachments to the program and sharing intimate secrets with it. He had not intended to create a therapist.

A survey conducted in 2023 found that a significant number of people said they preferred receiving feedback on their work from an AI rather than from their manager. The managers were not surveyed on their feelings about this.