On my fictional near-future Earth, the emergence of a networked artificial intelligence which can travel through both physical and wireless networks (copper telephone cables, Ethernet, WiFi, satellite signals, TV signals, even IR transmitters, optical transmitters, power cables and radio waves with the right equipment and enough patience) has created a separatist society of people who hide from it, attempting to evade its reach. As a result of fear, superstition, and alienation, this society have come to believe that any sort of network, however rudimentary, can be a pathway for the AI, even if in practice the AI can only travel through networks using the TCP/IP stack.*

This society are afraid of networks, however they define them, but they're not technophobic.

  • What sort of level of technology can I reasonably expect them to possess?
  • Is there a baseline decade in our technological development that their society would most resemble?
  • What kinds of technology, even if they don't appear networked, would they not have access to?

Some technological advancements I believe they would be able to possess include:

  • Standalone, airgapped computers. They might be afraid of these and see them as a huge risk, but they could use them if they needed to.
  • Various quality-of-life appliances like fridges, washing machines, and toasters.
  • Guns, flares, rocket launchers, and all 'dumb' weapons.
  • Cars (and even simple, vintage aircraft?) that have had their radio receivers (and entertainment systems) removed.
  • Farming and factory machinery.
  • Solar and wind power generators.
  • Batteries and capacitors, but perhaps not a power grid...

* Which is not to say that any form of communication cannot, with enough effort, be converted to use TCP/IP... even smoke signals. But the AI simply wouldn't bother.

  • 2
    $\begingroup$ Under a computer's hood, you can planely see that one chip is networked to all the others chips, looks dangerous to me, but a mains power-grid - there's none of those dangerous chip things, just light-bulbs and the fridge's compressor. There doesn't seem to be a clear demarcation point if aircraft (which, are flying computer networks that rely on satellite and ground-station signals to navigate) are allowed. Can you clarify where the line is? $\endgroup$ Jun 28 '21 at 11:42
  • 1
    $\begingroup$ @ARogueAnt. This seems like a very reasonably observation for someone to make, I agree! I suppose if someone was well-informed about computers, they personally could know that it was possible to keep one disconnected from all other networks, especially if it were battery powered. I agree about aircraft, I rather meant very simple, vintage single-wing craft. $\endgroup$ Jun 28 '21 at 12:00
  • 1
    $\begingroup$ I'm going to take a stab and say that the biggest barriers to advancement will be those personal networks that will need to compensate. Trade will be complicated (even if you allow something as simple as a point-to-point telegraph, ships won't have nav & Comm, and industry will be crippled accordingly) and individuals will need to carry books and knowledge physically. Tech will need LOTS of engineers. My solution was universal manufacturing devices, so every individual can make a replicator, then make high tech stuff. Civilization would collapse, though. $\endgroup$
    – DWKraus
    Jun 28 '21 at 13:08
  • 1
    $\begingroup$ Given that there were no computer networks until the late 1970s -- or at least, there were no computer networks in general use -- I would say that it is safe to assume late 1970s technology. Just watch some films from that period. They had aircraft, motor cars, fixed-line telephones, photocopiers, electric light, record players (and cassette players! and tape players!), they have television (both broadcast and closed-circuit)... In general they had everything we use today, except our smart "phones" and personal computers. $\endgroup$
    – AlexP
    Jun 28 '21 at 15:50
  • 1
    $\begingroup$ @AlexP Reread the question. It's about any transmission and network, not just the ones that can reasonably transmit the AI. Their society distrusts ALL networks, even radio transmissions and power grids. I didn't give an answer because it's REALLY open-ended; I just gave a guess. $\endgroup$
    – DWKraus
    Jun 28 '21 at 16:29

Computer networking as we know it started being developed around the 1960's. But remotely controlled machines (as in receiving commands from afar) go further back:

In 1894, the first example of wirelessly controlling at a distance was during a demonstration by the British physicist Oliver Lodge, in which he made use of a Branly's coherer to make a mirror galvanometer move a beam of light when an electromagnetic wave was artificially generated. In 1895, Jagadish Chandra Bose demonstrated radio waves by triggering a gun and sounding a bell using microwaves transmitted over a distance of 75 feet through intervening walls. Radio innovators Guglielmo Marconi and William Preece, at a demonstration on December 12, 1896, at Toynbee Hall made a bell ring by pushing a button in a box that was not connected by any wires. In 1897 a British engineer and professor in the Electrical Engineering Department at King's College London, Ernest Wilson, had invented a remote radio control of torpedoes and submarines that was controlled by "Hertzian" wave.

Since all it takes in order to have a network is electronics + remote control capabilities, and the harder requirement to sustain AI lies in the complexity of the hardware (as long as there is a machine to run it, somebody will eventually code it), then your people are mostly afraid of anything electronic and/or anything that can be operated remotely.

Notice that a society is very influenced by the media it consumes, even if it is not realistic. Marvel has this character, called the Vision, who is an android - a fully aware AI in a human-like body. He was created in 1968. It would take another 46 years for a chatbot to arguably pass the Turing test, and we are possibly decades away from a truly aware AI. AI's overtaking the internet has also been the stuff of superhero comics ever since the Internet became mainstream, yet even the IoT is still a growing thing. The fact that science fiction predicts things decades or centuries before they become mainstream might lead those people into becoming neoluddites. They will reject anything that might become a monstrous AI in literature.


Networks don't have to be electronic -- some early attempts at the telegraph involved making bubbles of different sizes float up in a water tank. A network is, basically, any means of transmitting information. A group of people talking to one another is broadcasting information wirelessly (sound) to multiple nodes (listeners), who process the information and translate it to actions or responses.

At the bottom level, computers are powered by diodes, which are tiny relays, mechanically actuated by tiny flows of electric current. One diode is hardly a network, and two diodes is like two mechanical switches -- and every mechanical switch transmits information in the same way that a diode does, by being either on or off.

So, for example, if a car has AC, headlights, and an engine, then that's three subsystems requiring at least three bits of information to run (three mechanical switches); and if the key has any complexity at all then it also requires several switches inside it. When turning on the engine, several other parts of the car also activate: power steering, breaks, whatever; that's one switch sending multiple "on"-signals. Every one of these signals is a bit (or multiple bits) of information.

Your society will have to answer three basic questions:

  1. How much information counts as "networked information" -- how many "on/off" relationships can one device send/receive with another device?
  2. How many devices may communicate with one another by means of these signals at one time, without direct human intervention?
  3. How many moving parts counts as "one device" for the purpose of determining communication between devices?

If the answer to #1 is "1 bit", then you could reasonably use any mechanical switch, but you can't do any logical functions (like a "j/k flip-flop"), so you can't have a house where two light-switches change the state of the same lightbulb.

If the answer to #2 is "2 devices", then you might have to flip 3 to 5 switches to get your car started.

If the answer to #3 is "one moving part", and #1 and #2 are "1 bit" and "2 devices", then your door-locks are limited to skeleton keys with only 2 stages, because one turn of the key can only actuate two devices, and each device is defined as one moving part (one tumbler), and each tumbler may only perform one action on the bolt. This is an extreme example, but I'm trying to establish consistency across technologies.

Another one: imagine a person presses a button which links a large battery to ground, through a coil, generating a magnetic signal. Across the room, an RLC circuit picks up the field, opens a diode through a local source of electricity, and powers another coil which drives a piston across your window, drawing the curtains. If each electric device counts as a small moving part, you have arguably six moving parts on one 1-bit signal:

  • button,
  • battery/coil
  • RLC circuit (minimum 2 devices excluding resistors: L and C)
  • diode
  • piston/coil

That can be construed as:

  • 1 human interaction, 1 device (6 moving parts), 1 bit of information being transmitted once from human-to-device
  • 1 human interaction, 2 devices (one with 2 moving parts, the other with 4 moving parts), one bit being transmitted once from human-to-device, and once device-to-device.
  • 1 human interaction, 6 devices (1 moving part each), one bit being transmitted once from human-to-device, and 5 times device-to-device.

So, that setup would be allowable if they said "one device has 6 moving parts, and only one device may actuate per human interaction", or if they said, "one device has only one moving part, and 6 devices can communicate with one another per human interaction".

Once you've got the rules defined, then each piece of technology will have to be reworked according to those rules.


No technology or social interaction.

You can pass on code and network with other humans directly, telling them machine code to repeat and recite. If they were afraid of all networks, they would be afraid of all human interaction and as such, would be afraid of all technology that required social interaction.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .