AI and its brief history

Artificial Intelligence (AI) is the branch of computer science that builds machines and programs able to perform tasks that normally need human intelligence, such as learning, reasoning, problem solving, understanding language and perceiving the world. John McCarthy, who coined the term in 1956, defined it as "the science and engineering of making intelligent machines".

Brief history of AI (chronological)

1. Gestation (1943-1955)

• 1943: McCulloch and Pitts proposed a model of artificial neurons.
• 1949: Hebb proposed a rule for learning in neurons.
• 1950: Alan Turing published "Computing Machinery and Intelligence" and proposed the Turing Test.
• 1951: Minsky and Edmonds built the first neural-network computer (SNARC).

2. Birth of AI (1956)

• Dartmouth Workshop (McCarthy, Minsky, Shannon, Rochester): the name "Artificial Intelligence" was adopted. Newell and Simon demonstrated the Logic Theorist.

3. Early enthusiasm (1956-1969)

• General Problem Solver (GPS, 1957), LISP (1958), Rosenblatt's Perceptron (1958), Samuel's checkers program, ELIZA chatbot (1966), Shakey the robot.
• 1969: Minsky and Papert showed the limits of single-layer perceptrons, and neural-network research slowed.

4. First AI winter (1970s)

• Programs did not scale (combinatorial explosion). The Lighthill report (1973) led to heavy cuts in funding.

5. Expert systems (1970s-1980s)

• Knowledge-based systems such as DENDRAL, MYCIN and XCON worked well in narrow domains. Japan's Fifth Generation project (1981) gave a further boost. Prolog and rule-based systems became popular.
• 1986: back-propagation was re-discovered and neural networks returned.

6. Second AI winter (late 1980s-early 1990s)

• Expert systems were costly to maintain and brittle; the LISP-machine market collapsed.

7. Intelligent agents and machine learning (1990s-2000s)

• Probabilistic reasoning, Bayesian networks, SVMs and data-driven learning replaced hand-written rules.
• 1997: IBM Deep Blue defeated world chess champion Garry Kasparov.
• 2005: Stanford's car won the DARPA Grand Challenge.

8. Deep learning and big data (2010-present)

• 2011: IBM Watson won Jeopardy!.
• 2012: AlexNet won ImageNet using deep convolutional networks.
• 2016: DeepMind's AlphaGo defeated Lee Sedol.
• 2017 onwards: Transformer models led to large language models and generative AI (GPT, BERT, image generators).

1943   1956    1969-79   1980s    1997    2012    2017+
 |      |        |         |       |       |       |
neuron  AI     1st AI   expert  Deep    deep   generative
model  born    winter   systems Blue  learning   AI