16

It seems that in recent years, when people talk about "AI", they are usually referring to neural net-based technologies (ChatGPT being the most well known example at the moment). But computer scientists have been researching artificial intelligence since long before modern neural networks were a thing.

What is the earliest computer technology to be referred to as "artificial intelligence" (by either a researcher, marketer, journalist, etc)?

Note that I am not asking for the earliest reference to AI, nor when the phrase was coined, since I believe people (such as Turing) were thinking about and predicting intelligent machines before computers had even got off the ground. I am asking for the first time someone said, "look, this existing technology is an example of artificial intelligence".

I am particularly interested in the exact phrase "artificial intelligence", but I am open to answers about synonymous phrases (including equivalent phrases in other languages).

T Hummus
  • 261
  • 1
  • 5
  • 10
    And we're still waiting for AI all these years later. The current wave of hype and nonsense from venture capitalists sure isn't it. – Alan B Aug 04 '23 at 08:29
  • 9
    @AlanB: When regular people finally figure out that ChatGPT is a language model rather than a knowledge model, hopefully they'll back off a bit. A lot of money will have disappeared by then. – paxdiablo Aug 04 '23 at 09:36
  • 2
    As soon as the term 'electronic brain' was coined. – dave Aug 04 '23 at 11:49
  • 8
    The "Mechanical Turk" of 1770 (https://en.wikipedia.org/wiki/Mechanical_Turk) was repeatedly denoted as "intelligent/sentient automaton" - even if it wasn't. (Automaton, that is. It proved to be intelligent, however) – tofro Aug 04 '23 at 14:58
  • I would really say this is a rather opinion based question - especially by asking about usage of an arbitrary term possibly in non scientific context. – Raffzahn Aug 04 '23 at 16:01
  • 3
    @Raffzahn a question about people's opinions is different than an opinion-based question. "XYZ was the first AI" is an opinion, but I didn't ask about that. "So-and-so was the first person to call something AI" is not an opinion, it's a fact (even though it's reporting on so-and-so's opinion) – T Hummus Aug 04 '23 at 17:21
  • @THummus No, not really. Asking for a generic recorded opinion is at best a random quibble . if not a semantic trick to get it injected. – Raffzahn Aug 04 '23 at 17:31
  • 3
    @Raffzahn I respectfully disagree. There is a long history of people using "AI" to refer to things that aren't actually all that intelligent. I think it's worth knowing when that started – T Hummus Aug 04 '23 at 17:57
  • 2
    @Raffzahn Admittedly, this is more a question about human perception of computing than about the computing itself. That perhaps makes it off-topic for this SE. But that does not make it opinion-based. I want objective reporting on people's perceptions, not people's opinions on what constitutes AI. – T Hummus Aug 04 '23 at 17:59
  • @THummus Reporting of perception is in itself based on perception thus opinion on that topic. Calling something AI is as much opinion than recording that opinion is. And that's the whole problem of that line of though. Random recording of random words. – Raffzahn Aug 04 '23 at 18:14
  • 3
    re: using AI to refer to things that aren't actually all that intelligent. A consensus on what isn't AI seems to require a consensus on what 'intelligence' actually is, and I don't think we have that. Sure, we can often tell when software uses cheap tricks to reach a conclusion, but we don't know that brain doesn't do the same. – dave Aug 04 '23 at 18:46
  • 4
    And that reminds me: "if our brains were simple enough to understand, we wouldn't be able to". – dave Aug 04 '23 at 20:36
  • When I was at university (1983-87) our AI lecturer was very clear that AI referred to a computer doing something that would require intelligence if being carried out by a human. – Frog Aug 05 '23 at 01:47
  • Interesting question! Wondered about the origin of the term "artificial intelligence" myself many times, and when it was first applied to describe a technology is also very interesting. – BipedalJoe Aug 07 '23 at 05:00

4 Answers4

17

Programs that demonstrate AI, or perhaps fake AI, were around in the 1960s. SHRDLU and Eliza have already been mentioned in other answers. Somewhat earlier, the checker playing program by Samuel has to be mentioned, because that program improved its strategy by studying existing recorded games between master human players. So it developed strategies that the programmer had not developed.

However, if the question is really about journalistic hype rather than anything real, the big bubble has to be in early November of 1952, when a Univac computer predicted the Eisenhower victory over Stevenson on live TV. This was an easy race to call, and many humans had called it. The software running in the Univac was nothing like artificial intelligence. It was just standard statistical analyisis. But the coverage was full of breathless predictions that office workers were all soon to be turned out of work, and machines were going to "take over" all intelligent work. I doubt if the phrase "artificial intelligence" was actually used, but the term "thinking machines" most likely was.

In 1956, the first serious conference about AI was held at Dartmouth college. The Lisp programming language was launched in about 1960. This was first language designed to facilitate research in this area.

https://home.dartmouth.edu/about/artificial-intelligence-ai-coined-dartmouth https://en.wikipedia.org/wiki/Lisp_(programming_language) https://history.computer.org/pioneers/samuel.html

Walter Mitty
  • 6,128
  • 17
  • 36
9

SHRDLU was a good example of a natural language processing tool from the late 60s, one which worked with a universe of blocks of various sizes, shapes and colors.

I remember seeing a demo where you could say (well, type) things like "take the biggest shape that's on top of a cube and place it next to a blue sphere".

The program used reasoning to evaluate and act on statements and, though clever, could still be tricked. It would also converse if it needed more information, or to answer queries.

Here's a snippet from the conversation on the Wikipedia page (linked above):

Person: Pick up a big red block.

Computer: OK.

Person: Grasp the pyramid.

Computer: I DON'T UNDERSTAND WHICH PYRAMID YOU MEAN.

Person (changing their mind): Find a block which is taller than the one you are holding and put it into the box.

Computer: BY "IT", I ASSUME YOU MEAN THE BLOCK WHICH IS TALLER THAN THE ONE I AM HOLDING.

Computer: OK.

Person: What does the box contain?

Computer: THE BLUE PYRAMID AND THE BLUE BLOCK.

Person: What is the pyramid supported by?

Computer: THE BOX.

Person: How many blocks are not in the box?

Computer: FOUR OF THEM.

Person: Is at least one of them narrower than the one which i told you to pick up?

Computer: YES, THE RED CUBE.

Person: Is it supported?

Computer: YES, BY THE TABLE.

Person: Can the table pick up blocks?

Computer: NO.

paxdiablo
  • 4,742
  • 20
  • 47
7

The most famous one was certainly "Eliza", developed 1966 by Joseph Weizenbaum. This has (technically) nothing to do with AI, but cleverly simulates AI using a thesaurus in a simulated therapeutical session (using this situation to its advantage, as therapeutical sessions following Carl Rogers approach will mirror the patient's answer in the next question, which is what Eliza does as well).

Weizenbaum himself was utterly shocked about the reactions to his program: Even renowned therapeutists started to talk about the possibility to completely replace "real" sessions with computer-generated ones.

I'm not sure if Eliza was ever denoted as "Artificial Intelligence" but it was certainly treated as such in the public. Weizenbaum himself used the keywords "heuristic programming and artificial intelligence" in his 1966 paper explaining the program (but not directly connected to Eliza), so the term must have been commonly known by then.

user3840170
  • 23,072
  • 4
  • 91
  • 150
tofro
  • 34,832
  • 4
  • 89
  • 170
  • 3
    And how does that make you feel? :-) – paxdiablo Aug 04 '23 at 09:27
  • 7
    @paxdiablo This is not about me. Tell me more about you – tofro Aug 04 '23 at 09:28
  • 1
    Touché :-) By the way, there is a recent "Advent of Computing" podcast covering Prolog (part 1), which also delves into Eliza a bit. That's one of my go-to podcasts for the drive to work – paxdiablo Aug 04 '23 at 09:38
  • 2
    Since I got a good grade for handing in a (very limited) natural-language processor for coursework for the AI module in my mid-1970s BSc degree, I don't see this as 'simulating' AI. Although I was personally simulating an AI programmer. :-) – dave Aug 04 '23 at 11:53
  • 1
    @another-dave That is, I think, disputable. Eliza lacks the "knowledge" part of AI and only has very rudimentary ML components (it does consider context for exactly one reply). Natural language processing is very limited, and it doesn't really "understand" what you're typing, but rather reacts on a very limited amount of keywords. Otherwise, it just rephrases your statements into questions. Not much "intelligence" here. – tofro Aug 04 '23 at 13:09
  • @tofro Wenn, I'd say everything up to today lacks any 'intelligence', not just Eliza. For the way the question is worded toward an outside notion of intelligente behaviour, Eliza does the job as good. – Raffzahn Aug 04 '23 at 16:38
  • 1
    Also in my degree course, we learned the cynical alternative definition: AI is whatever we don't know how to do yet. When we figure out how to do it, it's not AI. – dave Aug 04 '23 at 18:41
  • There was a course taught by Marvin Minsky named Heuristic Programming and Artificial Intelligence (or maybe the other way around) in the 1960s. – Walter Mitty Aug 05 '23 at 10:46
  • To those stating that Eliza has no knowledge, I would posit that even modern things like GPT also lack knowledge of specific subjects yet it appears intelligent to the untrained. In the same way we see animals in clouds (because those in the past that didn't have that imaginative ability, useful for shape detection in bushes, were probably eaten before reproducing), a lot of the work is done by the receiver of the information, often unbeknownst to them. Of course, it may even be that our intelligence is similarly unreal :-) – paxdiablo Aug 06 '23 at 22:21
  • 1
    @paxdiablo Well, that was probably the most important discovery that Eliza sprouted: That a lot of what makes up the impression of "intelligence" is in the receiption by the receiver. Thinking of it, I must admit I've seen that in natural intelligence as well..... – tofro Aug 07 '23 at 06:48
2

I guess one should not forget Chess.

While today classified as 'Weak AI' and seen as branch of 'Game AI', playing chess was in prior times always seen as a pinnacle of human capabilities and automated chess play thus a major proof of machine intelligence.

Within a limited scenario (3 piece end game) this was already practically shown by the Spanish El Ajedrecista, an electro-mechanical special purpose computer, in 1912. It received world wide fame - way before anyone coined the term KI or AI (*1).

Naturally chess was as well in after war world of emerging computers long seen as a major goal for archiving higher functions - not at least due it's clear defined nature enabling a definitive proof of achievement. A good overview how early attempts to archive such is layed out in Claude Shannons 1950 paper XXII. Programming a Computer for Playing Chess. A title that already points into the direction taken early on.

In this paper he ranks the achievements as follow:

This paper is concerned with the problem of constructing a computing routine or "program" for a modern general purpose computer which will enable it to play chess. Although perhaps of no practical importance, the question is of theoretical interest, and it is hoped that a satisfactory solution of this problem will act as a wedge in attacking other problems of a similar nature and of greater significance. Some possibilities in this direction are: -

  • (1) Machines for designing filters, equalizers, etc.
  • (2) Machines for designing relay and switching circuits.
  • (3)Machines which will handle routing of telephone calls based on the individual circumstances rather than by fixed patterns.
  • (4)Machines for performing symbolic (non-numerical) mathematical operations.
  • (5)Machines capable of translating from one language to another.
  • (6)Machines for making strategic decisions in simplified military operations.
  • (7)Machines capable of orchestrating a melody.
  • (8)Machines capable of logical deduction.

It is believed that all of these and many other devices of a similar nature are possible developments in the immediate future. The techniques developed for modern electronic and relay type computers make them not only theoretical possibilities, but in several cases worthy of serious consideration from the economic point of view.

One can't help noting how low language translation was rated. It does shed light on the complete different evaluation of complexity back then, doesn't it? Important for this question might be that Shannon did not write this as a mere abstract contemplation, but based on his knowledge of Computers already existing and in development.

While that paper was a strictly theoretical venue, not much after first programs to solve certain chess problems were developed. Theoretical, like Turing did as early as 1951, or more practical by Dietrich Prinz, who was a major contributor to the Mark 1, implementing the Mate-in-two problem as well in 1951. By 1957 a full chess has been implemented by IBM on the 704. The rest is improvement thereon.

The important point here is that all of those steps, from Quevedo chess player all the way to Deep Blue's win in 1997, were seen at that time as incredible signs of machine intelligence - and reported as such.

*1 - AI seems to be a direct translation of German Künstliche Intelligenz - as several early English reports of the late 1940s early 1950s on that topic do use the German term.

Raffzahn
  • 222,541
  • 22
  • 631
  • 918