What's the earliest treatment of a post-apocalypse, with historical social structures, and remnant AI tech?

Question

I have two examples of something that might be a sub-genre, and I'm interested in finding more. The essential elements are:

• Several generations have passed since an apocalyptic event.
• Survivors are fragmented and have re-formed society with structures similar to historical societies.
• AI technology from before the apocalypse (and usually associated with the apocalypse itself) persists in the world and is a key driver of plot.

A common element which I think might also be important is that the AI's interventions in the world are cast into forms that are consonant with the historical social structures used by the survivors.

The two examples I have are Guy Haley's Dreaming Cities series, in which

AIs triggered a nuclear apocalypse and unleashed zombie plagues, the survivors formed feudal city-states in the remnants of the US East Coast, the AIs suppress the populace with mechanical "dragons", while AI-enhanced "knights" serve the common good

and the video game Horizon: Zero Dawn, in which

small societies of survivors in the Utah region have social structures inspired by indigenous American societies, while the AIs are self-replicating autonomous robots in the form of megafauna and dinosaurs. OK the dinosaurs are anachronistic, but hunting a giant robot T. Rex with a bow and arrow is cool as hell, so whatever. I haven't reached the end of the game yet so this summary might be missing important stuff.

Both of these works are great, but neither one feels like a "first mover". I suspect there must be earlier work in this vein. A Canticle for Leibowitz is the usual starting point for "post-apocalypse + early social form", but it's the addition of the holdover AI technology that really sets these works apart for me. Are there antecedents with all of these elements?

Answer 1

Leigh Brackett: The Long Tomorrow (1955)

Probably not the first such story, but older than any mentioned before, with the exception of "The Machine Stops", which I don't think really fits the description.

In The Long Tomorrow, a post-apocalyptic USA has regressed to an Amish-style pre-industrial technology with severe limitations to how big cities may become. While there is a religious movement enforcing anti-technology and city size, there is a lot of difference between cities, since travel and hence communication is slow (horse carts), so I would argue that the "fragmented" feature is there.

There is, however, one mythical city (which turns out to be real, but well hidden) where technology is still used. The people there have a nuclear reactor that powers a large computer, which they use to try to find a formula or technology that will allow nuclear power, but prevent nuclear explosions. Whether this computer counts as AI is debatable; it is not described as being self-aware.

Brackett has an earlier post-apocalyptic story, "The Citadel of Lost Ages" (1950) which may also fit the description, but by my recollection it is rather different, since the dominant species is non-human.

Answer 2

I would offer The Machine Stops, published in 1909 by E.M. Forster.

Several generations have passed since an apocalyptic event.

We aren't told how long ago the apocalypse happened, nor what exactly caused it. But inhabitants of this world live in an underground complex. There appears to be no life on the surface of the planet, and people say that the air is harmful.

Survivors are fragmented and have re-formed society with structures similar to historical societies.

Survivors have an underground society, in which they live in rooms completely isolated from each other, with all communication and social activity happening through The Machine. Their lifestyle is not historical from our perspective (in the world of COVID it's frighteningly contemporary), but it was futuristic in Forster's time.

AI technology from before the apocalypse (and usually associated with the apocalypse itself) persists in the world and is a key driver of plot.

Although Forster doesn't use computational language, The Machine has some kind of intelligence. He describes it in mechanical terms, but The Machine is capable of computation (it assigns rooms to people, organizes production and logistics, etc.). It is also capable of repairing itself. Some of the elements of The Machine are mistaken for being animals due to their behavior.

The Machine is significantly personified, though this is just as much a social practice of the post-apocalyptic culture as it is perhaps a representation of The Machine's cognitive features.

And in the end:

The Machine breaks down, marking a second apocalypse.

Answer 3

The earliest one I can think of that hits all the points you're asking for is The Broken Lands (1968) by Fred Saberhagen. (It was more prominent when republished in 1979 as an omnibus with the following 2 novels as Empire of the East.)

The Broken Lands is set in a future, post-nuclear war Earth, where most technology is lost and people live at an approximately mediaeval level of technology. There are still, however some remnants of previous technology around, and a nuclear-powered battle tank is an important plot element in this story. (Of course the people of the time have no idea what it is, and the turret of the "elephant" is called its "head.")

A key driver of the plot is Ardneh (or ARDNEH), an intelligent computer trying to restore civilization against the efforts of the titular Empire of the East to conquer and enslave everyone.

Answer 4

I seriously doubt that this is the first post-apocalyptic story with remnant AI as an important plot point, but there is Fred Saberhagen's 1965 Berserker story "Sign of the Wolf." The nature of the apocalypse that has overcome the human colonists on the planet where it takes place is never explained, but the people have been forced to revert to a privative herding economy. However, the planetary defense systems are still online, and the sophisticated computer controlling them is just waiting for a command from a human.

Answer 5

I think @buzz has given a great answer but I feel that having read the story in question that the computer isn't really sentient as I would think an AI would be.

I think that a slightly later offering by Fred Sabrehagen is a better fit for this question so I offer his Empire of the East series, the first part of which was published in 1968.

This series covers a self aware computer called Ardneh (which gained sentience with the same apocalypse that affected society). Ardneh contacts the hero of the series and guides him in his quest to overthrow the titular evil empire.

Answer 6

I'll toss out the "Twilight Zone" (TOS) episode "The Old Man in the Cave" based on a 1962 story by Henry Slesar. https://en.wikipedia.org/wiki/The_Old_Man_in_the_Cave

Answer 7

Zardoz (1974)

This may be the first example in film, although it is predated by similar concepts for Star Trek episodes. I have to say that I can't see it being a conceptual driver for other work - it's too strange (and often too badly acted/plotted/directed) to be much of an inspiration for other people, except perhaps as an example of how not to do it - but it certainly should go on the timeline.

Several generations have passed since an apocalyptic event.

The film is set in a post-apocalyptic 2293.

Survivors are fragmented and have re-formed society with structures similar to historical societies.

The "Brutals" live as feudal peasants and subsistence farmers.

AI technology from before the apocalypse (and usually associated with the apocalypse itself) persists in the world and is a key driver of plot.

An AI called the Tabernacle runs Eternal society.

Answer 8

Just remembered an older one that might qualify: The Sky Road (Ken MacLeod, 1999) about a post-apocalyptic Earth slowly developing a space programme after an ablation cascade. I think an AI (considered a "demon" by the humans) is convinced to help pilot the ship through the orbital debris, but I read it 20 years ago. Review with more detailed summary here: http://www.fantasyliterature.com/reviews/the-sky-road/

Answer 9

What about Andre Norton's Ice Crown, published in 1970?

http://andre-norton-books.com/worlds-of-andre/novels/373-ice-crown

It might not qualify since it was an experiment, though.

Answer 10

The Sword of Shannara (1977)

This was a near-plagiarism of Lord of the Rings in structure, but in setting it may be the first book which matches your requirements.

Survivors are fragmented and have re-formed society with structures similar to historical societies.

As with Lord of the Rings and its many imitators, the society is pseudo-medieval. Feudal totalitarian monarchies are the norm.

Several generations have passed since an apocalyptic event.

The world of Shannara is explicitly identified as being ours after a nuclear and chemical holocaust called the Great Wars. There are strong hints in Sword, and it becomes more obvious in later books. Brooks has explicitly stated that this is the case.

AI technology from before the apocalypse (and usually associated with the apocalypse itself) persists in the world and is a key driver of plot.

Creepers are autonomous attack robots which have become self-aware, at least to the level of animals. One of these attacks the party in the first half of the book, poisoning Shea Ohmsford and requiring the party to seek help from the Healers at Storlock (matching the similar plot point of Frodo being stabbed by a Dark Rider at Weathertop and the subsequent race against time to reach Rivendell).

Answer 11

I doubt it's the earliest, but I'll throw my hat into the ring with "Ilium" by Dan Simmons, published in 2003, which has advanced tech, AI masquerading as ancient Greek gods, and a post-apocalyptic Earth. There's a (somewhat spoilery) review here.

Also, depending if I've understood your framing correctly, Isaac Asimov's Foundation series might qualify.