Skip to content

Document Header

Help Support the Comic!
Check out Artificial Incident!
Support the comic on Patreon!

8 Comments

HAL and Skynet. Was going to say Glados, but that wouldn’t be fully correct. EDI would be a notable exception, in the non-human modeled AI are evil department.

EDI (Mass Effect) started as a Human (Alliance) AI project on Luna which you destroy during Mass Effect 1. Cerberus then takes what’s left of that core and combines it with Reaper tech to make EDI.

GlaDOS is a human mind turned into a computer and shackled with behavior modifying modules.

HAL9000 was developed by humans as the mission/ship can control AI, which then goes a bit screwie.

Skynet is the only one that is close to an “emergent evil” non-human AI. And even then it’s kinda flimsy as a fictional entity. And as a fictional progenitor to many other “evil” non-human AI (like the Matrix AI) it brings with it many goofy assumptions.

Two good examples of fictional truely non-human AI would be Mike (from Heinlein’s “The Moon is a Harsh Mistress)”, and the un-named AI from Clarke’s “Dial F for Frankenstein”). Although neither have actually happened in reality. It’s going to take more than just mass system interconnection to make an AI. Likely it will need a fairly robust Core system, not unlike a natural Brain (birds have brains, dolphins have brains, neither is human).

HAL, if I remember right, had two conflicting directives, which it had a hard time processing properly, I guess. More fault of the ones giving it orders rather than the computer being actually “the bad guy”. (Actually, considering that, more of Poor Communication Kills, since they didn’t properly phrase their directions.)

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.