AI [Forum game players stay out]

[Lt Cmdr Matt]

My players please leave now! Spoilers below!


What exactly do we know about the Federation's attempts to build a self sustained artificial intelligence that is capable of being sentient, and a persona of a computer as seen in Andromeda?

For the game related part, in my campaign, I am thinking about having a "Beta" version of an AI on the Klingon Starbase, essentially it was stolen by the Romulan's and stolen again by the Klingons.

[Balok]

In TOS, we know that Kirk encountered humaniform robots in "What Are Little Girls Made Of?" The best of these, Ruk, was destroyed. But others existed, and so did the equipment for making them. No doubt "top men" are studying this equipment.

There was also the M-5, although one might argue that this was really a personality copied into a machine, and not an AI.

Humaniform robots were developed by an alien race, and seen in "I, Mudd". But these were not independent machines; they depended on "Norman" and his small control unit. "Norman" might well be considered an AI, but a rather crude once, since he was unable to think his way around a simple paradox. At the end of this episode, the robots were function again; this implies that Scotty was able to repair and reprogram "Norman". That suggests that, even if it couldn't build them, the Federation at least understand the basic operating principles. Of course, that could have been a simple hardware repair, after which the software would automatically begin executing again. Still, the Federation would certainly have made diagrams of Norman and a copy of his code as part of the repair effort. Top men probably have that, too.

There's Vaal, but that was probably damaged beyond easy repair during the Enterprise's attack. Besides which, there's no evidence it was any more aware than the Enterprise's own computer; everything it did could have been programmed responses to perceived stimuli.

In TNG, there were the Soong class androids like Data. It's unclear why, but Noonien Soong was supposedly the only man actually able to build one.

There were the Exocomps, which probably come the closest of anything so far, since they started out as simple machines and actually evolved an intellect.

Off the top of my head, I'm not coming up with any examples from DS9 or Voyager, but my Voyager knowledge is spotty.

Mostly, it seems that there's been a lot of trial and error, and that there could be a breakthrough at almost any time. The key to limiting it would be that software is complex, and the hardware it runs on requires special elements (for speed or power, or whatever reason).

[Sundowner]

Originally posted by Lt Cmdr Matt
My players please leave now! Spoilers below!


What exactly do we know about the Federation's attempts to build a self sustained artificial intelligence that is capable of being sentient, and a persona of a computer as seen in Andromeda?

For the game related part, in my campaign, I am thinking about having a "Beta" version of an AI on the Klingon Starbase, essentially it was stolen by the Romulan's and stolen again by the Klingons.

A Prime example is from VOY. The Doctor.

The federation has obviously created a program that can mimic human characteristis well enough to run a holo like the EMH series. Also, Dr Zimerman on Jupiter Station had his female holo friend running as an assistant, she displayed very human characteristics, such as concern for Lewis, even though he didn't like her attention, he just didn't turn round & reprogram her.

VOY also experimented with running the EMH as an ECH, so they can run the ships if required. Holo emmiters on all sections would be a given though if you want a holo to run like 'romy on the Andromeda.

Sundowner.

[Ergi]

Stolen by the Romulans, eh? You aren't planning to let the players trigger a certain xenophobic sub-routine?

-"Computer, activate sonic shower!"
-"Sonic shower acivated. 6 minutes, 12,000 Watts."

Edit: BTW, you know your site cannot be accessed? I just thought I should mention it because of your signature.

[Tobian]

I think the general state of Federation programming is such that they can generate forms of artificial intelegence. The Enterprise D computer was fairly sophisticated, and respondedin a generally 'inteligent' way, though still relativelly mechanistic. In general I'd say they started to move away from that because they were getting TOO inteligent, and if they are inteligent it brings in all sorts of rights issues. Point in case - Data and the EMH's matrix having to undergo rights hearings in Federation law to determine their status.

Other spontaneous inteligences include Moiriarty in TNG 9and his wife) - which is an interesting subplot for anyone wanting to find a data cube in the wreckage of the D and the residents of Fairhaven in VOY, which were more or less completelly aware of their enviromnent (their perceptual subroutines were malfunctioning); There's Minuet, which the Bynars programmed, though she may have just been a very sophisitcated piece of software rather than true artificial inteligence.

I think with the case of inteligence, the computers that the Federation have are fast and powerful enough to run artificial inteligence software by the mid - late 24th century - it's just a question of someone skilled enough to write all of the nuances into something which works - something which is completelly independant of it's programming and self aware - and that's the tricky part!

In the case of Data that was a sophisticated positronic brain and in the case of the EMH it was an adaptive heuristic database. Both of them had severe limitations in their software however. In the case of Data it was that he had a finite ability to 'learn' and reporgram himself. He was an adept mimic, but he could only copy, not emulate, not 'be' - he coul do an impression of someone who was upset, but couldn't become upset (untill his empotion chip).

In the case of the EMH his primary limitation was the size of his database and the fact he far exceeded his expected useage. I suspect that from the portrayals on screen that there was a finite ammount of sophistication any piece of software could have before the computer became overloaded, moreso than the continuin 'memory bufer' plotline - because effectivelly he ws being 'run' on a number of discreet processors within the computer core, and the physical lag effect of having his brain located about the ship would be a problem - creating a finite 'size' that his programme could be.

[Lt Cmdr Matt]

Does it sound feasble that the DTI is trying to develop artificial intelligence to interface with the humans and other life forms? I am talking in the 2370's and top secret at that.

[Tobian]

Hmm well as I seem to recall - in the TOS era, they had a psychotricorder, which could record the neural engrams of people and replay them (yes it was incredible how people always remembered themselves from the 'camera angle' rather than out of their own eyes ).

The Romulans were known to have technology (Poor Geordi La Forge, he was always TNG's O'Brien )which could read and manipulate people's memories, and the Dominion certainly could create entire scenario's in people's minds (such as when they first encountered the Dominion)

So I would say that they COULD, WHY they would do it is another matter. I can see Section 31 doing it, just because it would be an immenselly powerful way of manipulating people.

In a sense the 'artificial inteligence' subroutines, consisting of an adaptive hueristic matrix could 'learn' how to interpret a particular individuals thought processes and follow them, for example if you wanted a 'firefox' type scenario of a plane that could fire / fly according to it's users thoughts etc... However even in the 24th century I would say there would still need to be a 'learning' period while it got the hang of your way of thingking.

Probably the best example of something like that would be 'Alice' though of course she wasn't Starfleet made, the Technology that made her was on a comparable 'starfleet' level.. So I suggest you watch that episode again for the neccessary technobabble

Why, what kind of 'interface' were you thinking of? Direct mind interface, or more the pure Andromeda - because I think it could relativelly easilly do an Andromeda style interface.. The question would be on the ethical considerations of giving a machine the wholeistic controll of a Starship - Starfleet ships have always been 'hands on'.

But the answer is definatelly yes, because we have the EMH as a comparitive example.

[Lt Cmdr Matt]

I think it would be a neat plot arc to encounter this AI in the databanks of the computer on their station, and then have to deal with it. Maybe a bit too cliched.

I am just thinking a straight Andromeda scenario, not interested in the mind control yet, although that might be something for later.

[Tobian]

Don't get me wrong, you shouldn't let a constraint get in the way of a story ark... But knowing the ethical problems within the show helps add some fuel to your fire. If you can answer within your head why Starfleet would have thought it was a good idea (at the time) then that gives you motivation for your characters.

In Starfleet just deleting it or turning it off might not be as easy, because the theme of the show is often about 'rights' So it should create all sorts of headaches for your players !

[JonA]

Check out this thread:

http://forum.trek-rpg.net/showthread...ghlight=LCARSI
http://forum.trek-rpg.net/showthread...ghlight=LCARSI

And Killerwhale's Expanded Starship Operations guide - a new feature is the LCARSI system.

Not canon but way cool none-the-less.