"AI" has become a catch-all buzzword that's applied to almost any new software technology, so let's be careful to define terms. Most TL10/11 software (including Loroi versions) will include some features that are today included under the umbrella of "AI" such as machine learning and data-training-driven-content-generation. While these are powerful technologies whose applications can sometimes give the appearance of intelligence, they are not true artificial intelligence. ChatGPT can shit out working code or an essay on a requested subject (which may or may not be accurate), but it can't think or decide or understand anything fundamental about the universe outside of the data that it has specifically been trained on. It can't invent or create something that doesn't already exist, or solve a problem that someone else hasn't already taught it how to solve.Bamax wrote: ↑Mon Mar 31, 2025 9:01 amDo Loroi ship systems use AI? I am aware Arioch created this story before AI became more mainstream, so it is possible the Loroi don't fool with AI at all.
If they do use spaceship AI, please tell me they are more competent than CONRAD (the spaceship AI for the cartoon Alien Monday on youtube).
So I was watching Alien Monday and was awe struck how awful the spaceship's AI was at keeping the crew safe. It was as if it was trying to get them all killed, and nearly succeeded. My issues I have with the AI, having watched the entire video, are as follows:
1. If the AI knows a crew member has been in a section of the ship for THREE days when it has other tasks to do, that's the kind of thing you would think it should at least check the camera footage and then relay it ASAP to any other crew members. The surviving crew member had to prompt the AI to even tell her where the missing crew member was even though it knew all along.
2. Instead of waltzing into the unknown to check on a crew member who has not left a part of the ship for THREE days, try using a voice intercom first by asking him is he okay. If no response you check the camera feed. If you have no cameras there then shame on you ship designers. If you have no voice intercom triple shame on them. You would then have to go yourself to check it out, but at least you would realize that something is seriously wrong and potentially unsafe before you waltz right into it unprepared. Like seriously, if a crewman has not left a room for 3 days that is a serious red flag.
However do "true" sentient, self-aware, decision-making Artificial Intelligence systems exist at TL10/11? Yes.
Do the Loroi use these sentient, self-aware, decision-making Artificial Intelligence in their military ship systems? No.
The first reason why is that they don't trust machines to make decisions for them. The analogy that comes to mind is the gun aiming reticle that fighter aircraft have had since the end of WWII. We trust mechanical or software systems to make recommendations based on calculations that they can do far better than any human, but we depend on a human to make the decision to pull the trigger. Combat UAV's may sound like a great idea, but I'd rather trust my life to a human pilot than to a system designed by a software engineer somewhere. I've worked in software engineering for too long to trust software engineers with anybody's lives. The telepathic Loroi take this to another level, both because they are not tech-nerds and don't trust anything they can't read with their telepathy. They're not quite Butlerian jihadists from Dune, but they're not too far off.
The second reason why is that on a military vessel with several hundred crewmen, intelligent systems are simply not needed. You can expect that Loroi systems such as sensor monitor software or targeting software or the like will have systems which behave a bit like autonomous intelligences, but they will always have a living Loroi in command of that system. On a civilian merchant vessel like Nostromo that has less than a dozen crewmen to operate a huge vessel, a higher level of automation might make some sense, but even in that case you don't need self-aware AI to flag that crewmember has been missing for a certain length of time (hopefully a lot less than three days), nor do you need self-aware AI to make a call on the intercom, check a video camera or any other sensor to check on that crewman. So sure, there are expert systems that are monitoring Tempest's external and internal sensors, and they will inform the operators if they recognize patterns that they have been told are important. However, they may dismiss as noise something that they're not trained to look for, and there's not much that can be done about that until and unless you get into Historian-levels of technology; until then, having a living operator to manage smart software will almost always be better than giving a sentient AI control. And even in the case of the Historians, that's not the whole story.