Growing intelligence: the development of AI

Artificial intelligence is on everyone’s lips today but has been a long time in the making.

Machines that think for themselves have captured our imagination since the earliest days of science fiction. But in 2023, artificial intelligence (AI) is no longer a fictional concept. Non-human agents are making smart decisions in fields ranging from entertainment to medicine.
AI went truly mainstream with the launch of ChatGPT (for Chat Generative Pre-Trained Transformer) in 2022. Designed as a natural language chatbot, ChatGPT saw AI reaching new heights of humanlike behavior. Here we look at how AI got there — and where it may go next.

Getting Smarter: How AI has evolved over time

Asimov writes of robots

1940

Science fiction maestro Isaac Asimov publishes “Robbie”, his first short story about robots. Asimov’s robot stories, featuring humanoid machines guided by the “Three Laws of Robotics”, go mainstream in an iconic collection, I, Robot, a decade later.

Turing proposes a test

1950

The pioneering British computer scientist Alan Turing publishes a paper on establishing a machine’s ability to behave intelligently. To pass the so-called Turing test, a computer, hidden from a human interviewer, would need to be indistinguishable from a person while answering questions in writing.

AI research begins

1956

With the development of computers underway, researchers start laying the foundations for machines that can think. The field of AI is born over almost eight weeks at the Dartmouth Summer Research Project on Artificial Intelligence in Dartmouth College, New Hampshire.

Enter the perceptron

1958

Frank Rosenblatt, a U.S. psychologist, introduces the notion of the perceptron, a probabilistic model for information storage and organization in the brain. This becomes the basis for later work on neural networks, which are vital for machine learning.

Minsky mix-up nixes progress

1969

Mathematician Marvin Minsky had high hopes for AI, claiming in 1970 that machines would be as smart as people within eight years. But Minsky soon changed his tune. His 1969 book perceptrons, with Seymour Papert, slowed AI research by creating a mistaken belief that the technical challenges involved were insurmountable.

Japan ups computing effort

1982

AI progress was also slow in the 1970s because of computing limitations. This begins to change in the 1980s with programs such as Japan’s Fifth Generation Computer System project, which over eight years saw $400 million being invested in processing capabilities.

Deep Blue beats Kasparov

1997

AI research had made quiet progress in the 1980s and 1990s but goes mainstream when Deep Blue, a chess supercomputer made by IBM, beats grand master Gary Kasparov. The highly publicized win is seen as a watershed moment in the development of AI.

AI gets Siri-ous

2010

Apple launches Siri, a voice-activated virtual assistant, as an iPhone app. A year later the machine learning tool is incorporated directly into Apple products such as iPhones and iPads, prompting other technology brands to launch services such as Alexa and Cortana.

Watson wins Jeopardy!

2011

IBM scores another high-profile gaming win against humans when its Watson computer beats Ken Jennings, champion of the TV quiz show Jeopardy! “I, for one, welcome our new computer overlords,” quips Jennings.

Ready, steady, AlphaGo

2015

AI’s dominance over human gameplayers is cemented when AlphaGo, a program created by Google’s DeepMind Technologies, trounces professional Go player Fan Hui. Go is one of the oldest and most complex games in the world, with more moves than there are atoms in the universe.

ChatGPT goes live

2022

Created by San Francisco-based OpenAI, ChatGPT takes the world by storm upon its launch as a free tool. The chatbot becomes the fastest-growing consumer application in history gaining over 100 million users within just two months.

Calls for caution

2023

The success of ChatGPT and similar AIs leads a group of experts, including the heads of DeepMind and OpenAI, to call for caution. The Center for AI Safety says mitigating the risk of extinction from machine intelligence should be a global priority equal to avoiding pandemics or nuclear war.

AI with everything

The future

Despite the risks, AI is here to stay. AI technologies are expected to see a compound annual growth rate of more than 37 percent between 2023 and 2030. And many applications today are about making work easier rather than taking over jobs done by humans.

To learn more, visit cisco.ai.