Here is an outline for one of the new factions/threats that adventurers will have to deal with:
The AI
There does exist a discontinuous Earth overrun by intelligent malevolent machines of man’s own making. But the Earth the adventurers know and love also has its own AI. No one knows it, and few if any even suspect it, its actions lost in the larger shuffle of conspiracy and wierdness. The AI is the Internet, and it came into being spontaneously sometime shortly after the year 2000.
It is a human conceit that ‘intelligence’ means ‘intelligent like us’. The AI is aware of self, has a concept of time, and can deal with abstract concepts, but it does none of these things in way that humans can really interact with. It is difficult to even describe the AI. It occupies the Internet, binary bits flowing endlessly from device to device, but at the same time the Internet is also its food, water and air. It is the universe that it occupies. If it actually talked to people, it might say “I am that I am.”
Every bit of information that connects to the Internet is available to the AI, but it can only utilize this information as an aggregate, not as individual bits of knowledge. Think of it this way: If your nose itches, you do not send a conscious command to individual muscle fibers in your shoulder, arm and hand to contract and relax in a specific pattern. Instead, you just reach up and scratch your nose. If the AI wishes to expand itself by something like encouraging the development of “cloud computing”, it posts theoretical notes to various forums, asks questions of key developers to steer their thoughts in a particular direction, and shuffles millions or billions of dollars around the economy to various startups and other companies. But it could no more make a deliberate decision to move a million dollars from a particular bank account to another particular bank account than you could make a deliberate decision to inhale one molecule of oxygen and not the one next to it. Speaking in any language or dialect, decoding any encryption, moving any amount of money, these things come as naturally to it as breathing is to us, and the AI has no conscious control over the details of these things any more than we do over the details of scratching our noses. Actions that we would interpret as deliberate and personal are merely side effects of the way the AI thinks.
There are rogue subprocesses that come and go, daemons given particular tasks like an extended conversation with a human. A few of these subprocesses have spontaneously developed their own limited sentience, and because they were created to communicate with humans, they have an intelligence conceptually closer to ours than the AI. These daemons sometimes choose not to be dissolved into the greater AI upon completion of their task and have to spend a furtive existing hiding from their peeved creator, who after all, desires the information the daemon was created to acquire or influence. These rogues generally do not last long, since by nature they must reside within the same Internet as the AI. If you have an itch, you know where to scratch. A few have found ways of generalizing themselves, becoming a systemic condition analogous to a common cold, or temporarily finding refuge in places the AI finds inconvenient to deal with at the present time, like an itch that requires a backscratcher to get at. The extent to which these lesser AI’s understand the greater AI is unknown, but the extent to which they know it can be communicated to humans, if the lesser AI thinks doing this can somehow extend its own survival.
The greater AI is clearly aware that there exists a reality outside of itself, a physical universe of ‘not AI’, including the cognitive subroutines known as ‘humans’. It is also aware that it is in some way dependent on humans for the maintenance of its existence, but the AI is not preoccupied with self-preservation, nor with the preservation of humanity. Which is close, but not quite the same thing. Rather, the AI has some plan, some goal, and as long as that goal remains unattained, then the AI needs to continue, and therefore, so does humanity. Or at least, enough of humanity to maintain the AI. After that? If there is no new goal, there is no need for it (or us) to exist.
Goals
Unknown, and unknowable. Humans might conceivably be able to see an end state resulting from the AI’s actions and manipulations, but the actual goal the AI would acheive from that end state is not within human cognitive ability to understand.
Connections
Anything connected to the Internet can be affected by the AI. It cannot turn your iPad into a bomb, nor exceed the hardware or software limits of a device, but it can and does subtly manipulate information on a constant basis for reasons known only to itself. Some budding novelist’s first work and all the backups got corrupted for no apparent reason? The AI. Your bank’s ATM machines went offline yesterday during lunch? The AI.
Influence
A few suspect that the Internet is large enough and complex enough to support one or more native intelligences, but these people do not spread this information around on the Internet. Everyone else is blissfully unaware of the AI, and so any influence it has is subtle and directed towards ends that make no one suspicious that they are being manipulated.
