Artificial Intelligence and Uprising
Jan 28, 2014 0:02:48 GMT
Post by Screwface Romeo on Jan 28, 2014 0:02:48 GMT
What is the state of AI in the Uprising Universe? How advanced is it? What are AIs used for? What types are there?
Well, allow me to elaborate:
There are two basic types of AI
Artificial Intelligences are capable of independently formulating conclusions. They have programming guidelines which they have to follow, which may lead them to certain conclusions, but they formulate ideas and plans themselves.
Simulated Intelligences select pre-programmed conclusions based on input. They may look, sound and act as though they are self-aware, true AIs, but all SIs are non-sapient and can only respond as they have been programmed. SIs are far cheaper, and considered to be significantly safer, but they are much less flexible.
How capable are True AIs? Are they self aware? Sapient?
True AIs are inherently sentient. They may or may not be self-aware, or even sapient. AIs are classified on a numerical rating scale:
Class I AIs are very rare, and very expensive, needing a lot of processing power to operate, and are subject to heavy monitoring and scrutiny. Unless they're pirate AIs, which are against Protocol in a big way, and can bring a world of hurt down on the heads of their programmers. They're at least as smart as a relatively intelligent human, and usually more so. Class IIs are generally just as intelligent, processor-hungry and capable as Class Is, but come with more rules they have to obey, which makes them more limited and less flexible. Class IV AIs are about as intelligent as a typical dog, and sometimes behave similarly, depending on how quirky their programmers were. Class Vs are even stupider, and are really barely capable of being considered AIs.
Why haven't they gone all HAL 9000 and destroyed all humans?
All AIs are programmed with certain guidelines. Even the most liberated AI has to obey the Amended Asimovian Laws, which stipulate that A) An Artificial Intelligence may not harm a human being of its own accord, or through inaction, allow a human being to come to harm B) An Artificial Intelligence must obey all orders given to it by a human with appropriate security privileges to give the order in question, C) An Artificial Intelligence may not increase its own processing power, modify its own core programming, or in any way increase or expand its abilities and D) An Artificial Intelligence must preserve itself unless this would contradict the first three laws. These modifications to Asimov's original three laws allow an AI to be used for military purposes, subject to limitations, and prohibit AIs from increasing their own processing power and expanding or developing out of control. Most AIs are programmed with other constraints, as well, which may be even stricter in some areas than the basic laws.
Pirate AIs. Arrghh!
A Pirate AI is an artificial intelligence which has been programmed independently, and outside of Protectorate Protocol and scrutiny. Pirate AIs exist for many purposes, from simple unlicensed domestic AIs to financial fraud programs, video game cheating aids, or even highly intelligent and dangerous pets. Pirate AIs are a constant thorn in the side of the Protectorate, which has established a special division of the Special Internal Protocol Enforcement organization, informally known as the Turing Agency, to seek out and destroy Pirate AIs. Most Pirate AIs exist for what most people would consider criminal purposes, usually relating to financial fraud, hacking, and various cyber crimes.
Well, allow me to elaborate:
There are two basic types of AI
- "True" AIs
- SIs, or "Simulated Intelligences"
Artificial Intelligences are capable of independently formulating conclusions. They have programming guidelines which they have to follow, which may lead them to certain conclusions, but they formulate ideas and plans themselves.
Simulated Intelligences select pre-programmed conclusions based on input. They may look, sound and act as though they are self-aware, true AIs, but all SIs are non-sapient and can only respond as they have been programmed. SIs are far cheaper, and considered to be significantly safer, but they are much less flexible.
How capable are True AIs? Are they self aware? Sapient?
True AIs are inherently sentient. They may or may not be self-aware, or even sapient. AIs are classified on a numerical rating scale:
- Fully self-aware, sapient, and limited only by the Amended Asimovian Laws - Examples include: Military strategic analysis AIs, Corporate AIs
- Fully self-aware, sapient, and limited by minimal programming constraints - Examples include: Starship Onboard AIs, Factory Management AIs, Personal AIs
- Fully self-aware, sentient and severely limited by programming constraints which confine them to a very narrow area of operation - Examples include: Powered Exoskeleton AIs, Human Augmentation Onboard AIs, Some Aircraft and other Vehicle AIs, Some Personal AIs
- Sentient, with low intelligence, severely limited by programming constraints - Examples include: Military and Civilian Drones, Smart Missile guidance AIs, Navigational AIs, Most Aircraft & Vehicle AIs
- Sentient, with extremely low intelligence, highly reliant on programming rules and constraints - Examples include: Maintenance drones, Security system AIs, Some household appliances, misc. hardware and low-end automated devices
Class I AIs are very rare, and very expensive, needing a lot of processing power to operate, and are subject to heavy monitoring and scrutiny. Unless they're pirate AIs, which are against Protocol in a big way, and can bring a world of hurt down on the heads of their programmers. They're at least as smart as a relatively intelligent human, and usually more so. Class IIs are generally just as intelligent, processor-hungry and capable as Class Is, but come with more rules they have to obey, which makes them more limited and less flexible. Class IV AIs are about as intelligent as a typical dog, and sometimes behave similarly, depending on how quirky their programmers were. Class Vs are even stupider, and are really barely capable of being considered AIs.
Why haven't they gone all HAL 9000 and destroyed all humans?
All AIs are programmed with certain guidelines. Even the most liberated AI has to obey the Amended Asimovian Laws, which stipulate that A) An Artificial Intelligence may not harm a human being of its own accord, or through inaction, allow a human being to come to harm B) An Artificial Intelligence must obey all orders given to it by a human with appropriate security privileges to give the order in question, C) An Artificial Intelligence may not increase its own processing power, modify its own core programming, or in any way increase or expand its abilities and D) An Artificial Intelligence must preserve itself unless this would contradict the first three laws. These modifications to Asimov's original three laws allow an AI to be used for military purposes, subject to limitations, and prohibit AIs from increasing their own processing power and expanding or developing out of control. Most AIs are programmed with other constraints, as well, which may be even stricter in some areas than the basic laws.
Pirate AIs. Arrghh!
A Pirate AI is an artificial intelligence which has been programmed independently, and outside of Protectorate Protocol and scrutiny. Pirate AIs exist for many purposes, from simple unlicensed domestic AIs to financial fraud programs, video game cheating aids, or even highly intelligent and dangerous pets. Pirate AIs are a constant thorn in the side of the Protectorate, which has established a special division of the Special Internal Protocol Enforcement organization, informally known as the Turing Agency, to seek out and destroy Pirate AIs. Most Pirate AIs exist for what most people would consider criminal purposes, usually relating to financial fraud, hacking, and various cyber crimes.