Android

Larsen · Unread post by **Larsen** » Thu Jul 08, 2004 4:02 am

Hey just wondering if anyone has ever played an android and how it went? Wondering because i'm reading the book and just not getting it i guess. the difference between standard artifical intelligence and advanced AI is what im talking about. I mean i understand advanced ai learns and standard doesn't but my question is more at lvl 1 what makes the two types different from each other?

Snowtiger · Unread post by **Snowtiger** » Thu Jul 08, 2004 8:31 am

A standard AI can only respond as programmed, while an advanced AI functions like a real person, taking in data and reaching conclusions on its own, consider these examples, assuming an android is given a small apartment in the research facility it was created in:

Standard AI:
Human: "How do you like it?"
AI: "That does not compute. Please rephrase query."

Advanced AI:
Human: "How do you like it?"
AI(after taking a look around the room): "These accommodations are acceptable. I presume I can freely express my needs if something needs to be changed?"
Human: "Yes, just ask me if you need anything."

Toasty_Duck · Unread post by **Toasty_Duck** » Thu Jul 08, 2004 9:14 am

android: Think Data from STNG
robot: Think of those carpet cleaning things, or those robot lawn mowers

KillWatch · Unread post by **KillWatch** » Thu Jul 08, 2004 1:10 pm

for the AI to know what the person was referring to when he said IT it would have had to have had previous experience in the situation or at least been programmed with it.
If not then the AI would then ask to rephrase or to define IT.
once explained the AI can then use the reference in the future
it learned
however it would not LIKE anything. the room would be appropriate or acceptabble. is there a place where he can charge his batteries or refuel? tools to make repairs? the robot despite AI would only require necessities and everything else is frivolous. As long as there was enough room and suitable equipment he could stay in a closet

the basic AI, unless programmed with a slang dictionary would only be able to speak/understand proper english or try to correct you in your obvious mishandleing of the langauge. If a map to a location is 10 years old and a door has been walled off the robot would be lost. A kind GM might let him use his sensors to detect how thick the wall is and if there is anything on the other side
a cruel gm or just one thats pissed off would cause a system error crash

Larsen · Unread post by **Larsen** » Thu Jul 08, 2004 2:07 pm

first off lol mr montague at the system erroir crash. but ok i see the difference now but how easily do they learn? is it more like us ie=we need constant reminders til it sinks in or is it a one time thing where you teach it once and he remembers for all time? also the book i have has optional personality thing is there more options in other books?

Snowtiger · Unread post by **Snowtiger** » Thu Jul 08, 2004 2:30 pm

I think once an advanced AI has learned something, it will be stored in its memory for future reference, it wouldn't have the sometimes annoying human trait of forgetting something, so if it made an error the first time it tried something, it would know that the method that led to the error was flawed and it would try again by using another methods until it succeeds in the given task.

For example if an assassin android with the advanced AI would be unsuccesful in the termination of its primary target although all secondary objectives were successful, it would know that if it altered the process a bit, it would perhaps be succesful in terminating its primary target(like deciding to wait until the target would be alone until getting close enough to terminate the target).

KillWatch · Unread post by **KillWatch** » Thu Jul 08, 2004 2:43 pm

too many variables
like would he consider fighting everyone so that his target can be eliminated
would he decide to go all out and destroy primary target not taking any self preservation
would he try to correct the actions of his first encounter to combat the resistance?
or like you said just wait til later

acreRake · Unread post by **acreRake** » Sat Jul 10, 2004 6:16 pm

prophet118 wrote:personally id like to see the standard AI removed... no one likes to play that, now as a GM option for an NPC, thats fine.... or for rules when your hardware guys want to bust out an android, thats fine.

Awe, c'mon. There are lots of possible uses, checkout the topic about the robot vehicle with the back up AI (kinda like the Batplane... or whatever that thing is called) Also when building multiple robots with your budget, especially "remote probes." Or specially designed mini-robots that are programmed to be able to repair the main one... stuff like that.

ZEN · Unread post by **ZEN** » Mon Jul 12, 2004 10:05 am

When opting to play an AI character, I look very carefully at my skill selections, while I'm playing this character I would want a fair degree of human-like behavior, all pre-programmed, automatic response sort of stuff, such as the Terminator machine/flesh cyborg CPU. Just because it's an AI doesn't mean it is stupid, it must have pretty good running software and some fuzzy logic in order to just move around at a decent speed in an ever changing environment.
While an advanced AI is more human like in the way it can adapt and learn faster, the normal AI may also be able to call on remote databanks in order to adapt to new situations.. Matrix style!
Say the AI is carrying out it's mission to secure an area and then install a camera system that will cover the most ground while keeping it out of sight.. no problem, that's complex and seems like a creative task, but it all falls into it's programming and while it does the job, it actually has more CPU power available to it than an Advanced AI, which has to think about it more and actually is fairly creative.
Then something unexpected happens and the robot has to react to an intruder.. the Advanced AI reacts and comes up with a plan based on the current conditions and what it aims to achieve. The AI calls up the suitable program, but if it is not equiped with combat software, perhaps it just retreats and dials home for a quick update? And while the combat program may be less flexible, it may be much better for that, covering situations and probabilities that the Advanced AI hadn't thought of.

Food for thought anyway.
8-)

acreRake · Unread post by **acreRake** » Mon Jul 12, 2004 6:17 pm

It might be fun if the personality program is considered to be extremely detailed. I'm not sure if i can explain this right, but: the "personality" of the robot could have all kinds of contingency sub-routines, a thousand ways of saying (or showing) that it can't deal with a situation, and default (and random) responses to all kinds of issues and eventualities. Especially if the AI is based off of an actual person the programmer is likely to include all kinds of esoteric knowledge, and reactions that you wouldn't normally attribute to and AI.

IE: i think a standard AI robot could be playable if the personality program was in depth and complex enough. Eh?

Snowtiger · Unread post by **Snowtiger** » Mon Jul 12, 2004 7:21 pm

They do allow players to play droids in SWRPG(which have to be programmed for everything they do if they don't have a Heuristic Processor, which facilitates the learning capabilities), so why should we not allow a Player to play a robotic character with Standard AI programming, if he wants to do so in the first place. Sure, it can be a little rough for the character to understand abstract concepts and such, but if the player wants to play a character like that, who are we to say it cant be done.
Sure, if you want to set up a house rule that insists on not playing characters like that, you can do it, but why not try it out first, see if its worth the trouble, and if it isn't... well, you know the rest.

And why not put up a house rule option, that a standard AI can be upgraded into an Advanced one if the robot gets mangled and the owner of the robot wishes it and has the funding and the equipment to do so(this would at least require a recent copy of the AI's core programming, a new body with enough processing power to support an advanced AI and a robot workshop of some kind)...

Snowtiger · Unread post by **Snowtiger** » Mon Jul 12, 2004 7:38 pm

prophet118 wrote:im not saying dont let people do it, im saying that its not exactly what i call fun...

as for star wars rpg... depending on which you are referring to, west end games version, the AI isnt a big issue..besides, they have yet to show me a droid in the star wars universe that doesnt have more personality than the other actors

I wasn't pointing fingers in any direction when I said that, AFAIK. I was referring to the D20 SWRPG...

acreRake · Unread post by **acreRake** » Tue Jul 13, 2004 3:21 pm

Snowtiger wrote:And why not put up a house rule option, that a standard AI can be upgraded into an Advanced one if the robot gets mangled and the owner of the robot wishes it and has the funding and the equipment to do so(this would at least require a recent copy of the AI's core programming, a new body with enough processing power to support an advanced AI and a robot workshop of some kind)...

That reminds me of a character from the story "Mr Boy" by James Patrick Kelly. The title character is an extremely rich kid in the near future who owns an android. Apparently all domestic androids in this world have "standard AIs" but he had a "chop-job" done "in the city" and is now as human-like as the other characters (who, incidentally are at various stages of losing their own humanity). I don't think a new body is necessary, certainly the cost of a brand new Adv. AI, though.

KillWatch · Unread post by **KillWatch** » Tue Jul 13, 2004 3:30 pm

here is a thought for the secondary AI. If you were to set up a data bank for information retrieval you would have to spend money on even more AIs to disperse the information and for the varied range of topics/skills

ZEN · Unread post by **ZEN** » Tue Jul 13, 2004 10:42 pm

Ha ha.. standard AI personality programming.. I don't think there would be a standard package. AI's modelled on famous people would be fun and probably quite common (depends on the programmer/s).
Instead of "Does not compute" (lets face it, a programmer would have to be a complete moron to include that phrase), the AI responds by going "Hmmm, I have no idea, what do you suggest?" or "I'm sorry, I'm new to this kind of thing", different responses based on the situation, modelled around what a personality would have normally said or done.
What would a dottering professor do? A professional security guard? A police officer, nurse or factory worker?
Best example of the response of an AI to a situation out of it's programming range is the scene from Star Wars episode one, were the Jedi walks up to the droid soldiers and states that he is taking a group of prisoners to another planet..
"Were are you taking them?" says the droid..
"To Coruscant" replies Jedi..
Droid thinks for a second, mumbles and hesitates, then snaps to the default answer..
"You're under arrest".
8-)

Roger roger

Larsen · Unread post by **Larsen** » Wed Jul 14, 2004 2:55 am

btw can a robot have an alignment if all they are programmed to do is follow their directives?

KillWatch · Unread post by **KillWatch** » Wed Jul 14, 2004 9:42 am

evil is a choice the basic AI doessn't have any. The advanced AI would however

Larsen · Unread post by **Larsen** » Thu Jul 15, 2004 12:33 am

that was more my actual question montague. because lets say a basic ai is programed to always obey john doe and must do it. well if john doe says kill all the children in this church basic ai must obey. so thanks montague you seem to have the answer to most of my questions lately.

ZEN · Unread post by **ZEN** » Thu Jul 15, 2004 1:40 am

An AI will have a basic set of prime directives that put limitations on it's programmed behavior.. the three laws of robotics are the best example of this..
A robot may not injure a human being or, through inaction, allow a human being to come to harm.

A robot must obey orders given it by human beings except where such orders would conflict with the First Law.

A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

Modification of the first law allows some variation in alignment.. after all, injure is pretty inclusive (even a tiny scratch is an injury, or what about stress and mental harm?).. maybe that could be changed to kill or severely injure, or permanently injure? Perhaps the "Through inaction allow a human being to come to harm".
There is no reason why a robot with this programming could not kill someone's pets, damage their property or so on, as long as nobody told them not to do it.
The second law can be almost entirely removed, or modified so that only specific individuals can order the robot around.
The third law may be bumped up to out rank the second or even the first law.. survival at all costs could result in actions we would deem evil.. but there is the answer to your question..

Robots do not have emotions, they are programmed, they see everything as either yes or no, and all questions are answered with logic and calculation.. they don't instinctively know anything or feel anything, they don't understand cuteness unless they are programmed to act in certain ways the programmer has dictated for the robot to react to cuteness.
The robot does not have an alignment.. the programmer does, and it is the programming that determines how the robot is going to react.. so, in game terms, while the player of the robot can choose the machines alignment, that player has to stick to that alignment's do's and don't's like glue, unless the programming is changed, that robot is going to be exactly what the alignment restrictions state, not just acting in a manner which follows the guidelines of that alignment.
8-)

Larsen · Unread post by **Larsen** » Thu Jul 15, 2004 1:58 am

so are advanced AI start like this? example robocop (i know its not a good example hes a cyborg) what i mean by this example was he had a set of directives that he could not break but they did not govern the way he acted they only restricted him from doing certain things. or is an AI more like terminator where they are programmed to be a certain way and then they just learn from there and never change their alignment?

KillWatch · Unread post by **KillWatch** » Thu Jul 15, 2004 9:41 am

I think it's both in the case of the advvanced AI

The AI would start out like the Terminator, but as it learns and grows and begins to judge things for himself it goes from the directives being his guidbook on how to proceed to things he is limited by

kind of like being a kid and your mom tells you not to play with the stove
that is a directive as you grow and learn your programming is modified to understand that playing with the stove could cause destruction and it's not the whole stove that is dangerous but the oven and the areas of fire on the stove itself

ZEN · Unread post by **ZEN** » Thu Jul 15, 2004 10:40 am

Nice example Mr. Montague.
The AI can learn, and any artificial mind that is that complex is going to have some chaos involved in it's mathematics, so yeah, you have to include the capacity for change in it's programming.. a bit of slack, so to speak, to stop it grinding to a halt over every little thing that it has trouble dealing with. Every directive, as simple as they sound in text, is incredibly complex and covers huge wads of data on every situation applicable, or sets of rules that apply to groups of conditions.. the real nature of the AI program is fictional, it doesn't exist yet, so as always, let your imagination wander, pluck an image that suits and run with it.
8-)

Borast · Unread post by **Borast** » Thu Jul 15, 2004 11:06 am

ZEN wrote:A robot may not injure a human being or, through inaction, allow a human being to come to harm.

A robot must obey orders given it by human beings except where such orders would conflict with the First Law.

A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

I once read an Asimov short story about a minimg colony on a Jovian(?) moon populated with hideously expensive androids and reasonable cheap human labourers/supervisers/etc. They found that the first law was costing them too much in replacement parts, because the 'droids would rush into hazardous situations (ie: radiation) to drag-out humans and fry their own more vulnerable neural nets. So, the next batch they bought had the "inaction" part removed from law #1. The droids were now capable of murder - lift a heavy weight, hold it over an unsuspecting human, and let go...gravity killed the human, not the 'droid... :twisted:

Sir_Spirit · Unread post by **Sir_Spirit** » Mon Aug 15, 2005 11:27 am

I can imagine that since robots are cheaper than humans is might be feasable to put them in middle management type positions or in control of a group of contractually indebted servants to a company.
Thus making your boss a robot.......

Jimmy Crat · Unread post by **Jimmy Crat** » Tue Aug 16, 2005 3:54 am

Like the girl in Bubblegum Crisis 2040.
:lol:

Sir_Spirit · Unread post by **Sir_Spirit** » Tue Sep 06, 2005 9:42 pm

Shadow wrote:Did I miss it or do robots not get a PP? I looked a couple of times and could not find it. I know that Bionics have to buy their PP, but I'm not seeing it for robots

Let me check....

Sir_Spirit · Unread post by **Sir_Spirit** » Tue Sep 06, 2005 9:58 pm

ACtually you only have to buy PP, if you want to get it UP and AREN'T using the cyborg conversion options.
Other wise it's pretty much set at 22(if you use the CCP) or 10( if just buying the stantdard hand/arms) form what I can tell.

It doesn't say anything about a robots pp in the Robots section so you could jsut assume it stats at 10 and has to be bought up like the Bionics or work something out with the GM. (Maybe just do a random roll if playing an android...)

CHeck the FAQ!

Unread post by **NMI** » Tue Sep 06, 2005 11:07 pm

Question: How is a robot's P.P. determined? I can't seem to find it in the book. Same thing goes for M.A.

Answer: There is a discussion of robot attributes with suggestions for purchasing P.P. in the HU2 GM's Guide on pages 12-14. [FYI, the cost of increasing the P.P. attribute is the same for robots and cyborgs -- TD]

znbrtn · Unread post by **znbrtn** » Wed Sep 07, 2005 11:44 am

Shadow wrote:Thank you for that info.

Has anyone else had difficulty making an android even with the max budget allowed? I started to make on last night and I bought .........

frame (reinforced), basic arms & legs, micro-fusion power, adv. AI, 5 attacks..........then I was out of money. Nothing else could be bought. It was discouraging to say the least.

yes, it's really hard to get precisely what you want with the robotics category most of the time. i've tried in the past to make a fifty foot tall robot(just to see if i could), and i was over budget before i even got to the legs. i really think they should give the category an overhaul, but i doubt that'll happen any time soon....

Tinker Dragoon · Sat Dec 31, 2005 11:29 pm

Shadow wrote:Thank you for that info.

Has anyone else had difficulty making an android even with the max budget allowed? I started to make on last night and I bought .........

frame (reinforced), basic arms & legs, micro-fusion power, adv. AI, 5 attacks..........then I was out of money. Nothing else could be bought. It was discouraging to say the least.

Yeah, this was addressed in the HUGMG. Apparently in the design phase, the cost of several key systems were increased without a corresponding increase in character budget. The book recommends either increasing each budget level by 1D4 million dollars or reducing the cost of fusion power systems by "a 1.5. dollars" (presumably this is supposed to be 1.5 million dollars).

Sentinel · Unread post by **Sentinel** » Sun Jan 01, 2006 2:22 pm

I've always enjoyed playing artificial beings, both true robots and androids. There are different response types, or personalities, that can be played, from the very human-like, to the literal minded automaton.
I never really had much difficulty with the budgets for player characters, and villain/NPC robots and androids could be built in stages where the criminals/villains could steal more money, or borrow it from brokerages or some other group.

Gomen_Nagai · Unread post by **Gomen_Nagai** » Sun Jan 01, 2006 4:23 pm

if you want a high tech world I'd just use the space craft budgets from AUGG and apply to the Robotics category.

I read there was a Genetics category that had a similar insane budget.

The other option is to make Giant Combustion Engine robots ala Sakura wars.
but in any case, the HU robotics section all requires brand new options and budgets.

Snowtiger · Unread post by **Snowtiger** » Sun Jan 01, 2006 6:38 pm

gadrin wrote:I guess the HUGMG means subtract 1.5 million from the m-fusion.

Just checked that, it says subtract the 1.5 million from all fusion power plants, not just the micro-fusion plant. You could still add the suggested 1d4 million to the budgets, just to be sure that you've got enough to finish the char.

Sentinel · Unread post by **Sentinel** » Tue Jan 03, 2006 1:24 pm

oh man! Robotic PS in HU is much more powerful than the pitiful stuff in Rifts. actually it's not that bad, but a robotic PS of 20 in HU is incredible compared to its Rifts counterpart. Whew !

Yes, in general.
Supernatural PS in Rifts enables a character or creature to liftt 100X its' PS in pounds. This is half as strong as EXT PS in HU.
Our Supernatural PS allows one to lift 500X PS in pounds: no Rifts character possesses that level of strength.

Sentinel · Unread post by **Sentinel** » Tue Jan 03, 2006 2:15 pm

gadrin wrote:
Sentinel wrote:
oh man! Robotic PS in HU is much more powerful than the pitiful stuff in Rifts. actually it's not that bad, but a robotic PS of 20 in HU is incredible compared to its Rifts counterpart. Whew !

Yes, in general.
Supernatural PS in Rifts enables a character or creature to liftt 100X its' PS in pounds. This is half as strong as EXT PS in HU.
Our Supernatural PS allows one to lift 500X PS in pounds: no Rifts character possesses that level of strength.

yeah I was always confused by the pictures of Rift's Crazys lifting a tank or even taking down a juicer. I suppose the odd* crazy could but for the most part, it seems like just hype.

* very odd

I "fixed" PS in my campaigns for the most part, when I switched to straight SDC. HUII became the standard for PS, and I fit the other games' PS ratings to HU.
Brute PS in AtB, and Bionic/Enhanced PS in Rifts became EXT PS.
Brute PS in AtB and Robotic PS in Rifts became Superhuman.
Crushing PS became Supernatural PS.
Rifts characters who can lift 100X get "upgraded to EXT PS.
I then take supernatural creatures and creatures of magic and fit their PS on a case-by-case basis.