Bay 12 Games Forum

Please login or register.

Login with username, password and session length
Advanced search  
Pages: 1 ... 21 22 [23] 24 25 ... 29

Author Topic: Reactions to Chinese White Balloon of AI Awareness  (Read 24960 times)

jipehog

  • Bay Watcher
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #330 on: March 10, 2023, 11:09:10 pm »

A very nice conversation: Debate: Will ChatGPT Do More Harm Than Good? with Gary Marcus and Keith Teare

----

Also a follow up about the legality of AI search engines (and by extent content generation as whole): Artificial intelligence is reaching behind newspaper paywalls
Spoiler (click to show/hide)

Keep in mind this part of bigger trend where the EU and other countries passed copyright rules that force US tech giants to licensing content from local providers instead of exploiting their monopoly status to use it for free.
« Last Edit: March 10, 2023, 11:10:51 pm by jipehog »
Logged

King Zultan

  • Bay Watcher
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #331 on: March 11, 2023, 03:01:35 am »

That poor AI is to dumb to realize that it's plagiarizing stuff.
Logged
The Lawyer opens a briefcase. It's full of lemons, the justice fruit only lawyers may touch.
Make sure not to step on any errant blood stains before we find our LIFE EXTINGUSHER.
but anyway, if you'll excuse me, I need to commit sebbaku.
Quote from: Leodanny
Can I have the sword when you’re done?

anewaname

  • Bay Watcher
  • The mattock... My choice for problem solving.
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #332 on: March 11, 2023, 11:41:52 am »

Who is dumber? The AI that bypassed the GUI-blocks, or the coder that sent the entire datablob and relied on the GUI-blocks to restrict the user from the paywall'ed portion of that data.
Logged
Quote from: dragdeler
There is something to be said about, if the stakes are as high, maybe reconsider your certitudes. One has to be aggressively allistic to feel entitled to be able to trust. But it won't happen to me, my bit doesn't count etc etc... Just saying, after my recent experiences I couldn't trust the public if I wanted to. People got their risk assessment neurons rotten and replaced with game theory. Folks walk around like fat turkeys taunting the world to slaughter them.

EuchreJack

  • Bay Watcher
  • Lord of Norderland - Lv 20 SKOOKUM ROC
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #333 on: March 11, 2023, 08:52:18 pm »

That poor AI is to dumb to realize that it's plagiarizing stuff.
Hm, "dumb" isn't the word I was thinking.

So the risk isn't Skynet setting off all the World's Nukes.  The risk is Skynet revealing all the World's Nukes.

Eric Blank

  • Bay Watcher
  • *Remain calm*
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #334 on: March 11, 2023, 11:06:19 pm »

Sounds like the risk is the holders of the world's nukes not ensuring Skynet is prevented from finding and divulging the locations of the world's nukes. It's reasonable right now for website developers to prevent ai using their site through whatever means, and whitelisting ai that are proven to be programmed not to break the law.
Logged
I make Spellcrafts!
I have no idea where anything is. I have no idea what anything does. This is not merely a madhouse designed by a madman, but a madhouse designed by many madmen, each with an intense hatred for the previous madman's unique flavour of madness.

Starver

  • Bay Watcher
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #335 on: March 12, 2023, 07:00:04 am »

Difficult one, that. Implanting the "½th Rule Of Robotics" of "...must obey all relevent laws of copyright and IP[1]".

The whole Paperclip Maximizer concept is focussed at scraping and spewing information (though the ChatGPT was given precurated scrapes, to some degree, obviously proper search engines are going to have ongoing scraping duties[2]), and who is to say how we can add safeguards that keep up with the whole system that is supposed to be hands-off, self-developing, focussed on the basic premise of information flow. If it comes across a megadump of usernames and passwords, perhaps that's explicitly barred from publication but (for an AGI, that can reasonably discerne the usefulness) still is internally useful in accessing whole new clouds of private but not explicitly barred information, in ways the developers never appreciated. That's one scenario, and perhaps (because it is now thought of), it can also be barred from being a viable aim of a hypothetical 'webmind', in advance, but the mythical perfection of Asimovian 'perfect control' over an AI[3] is a pipe-dream, and that's without V'GER levels of future developments suddenly presenting us with far more problems than mere information tsunamis...


(Probably won't go that way, but you can't rule out that it will if you don't even know how far it could go. Even if I'm not confident that AGIs of sufficient complexity are imminent, a 'dumb AI' can still go wrong in various ways (through emergent behaviour/capabilities) that might or might not be foreseen (or encouraged) by its human creators/caretakers/interactees. As we start to see, with the Paywall scenarios. Though I suspect I'm more a chearleader of that 'problem' than I perhaps should be, anyway.)



[1] "... except where such orders would conflict with the Zeroeth Law".  Because if we're capable of doing the other, to over-rule the 1st-3rd, best to also implement the zeroeth, which might even be the prime solution to not revealing nuclear launch codes to random emotional teenagers, and the rest, as well.

[2] I'd actually hope that this removes the trend for Search Engine Optimisation through spam, as AI-optimised Search Engines learn to better filter out every trick that mere ~~~~BUY LEATHER GOODS, VIAGRA, W0RLD CUP TICKETS HERE!!!~~~~ humans try with their illegitimate attempts to sway the 'dumb' webcrawlers.

[3] Which was, anyway, proven (in-universe) to be imperfect. Time and time again...
Logged

jipehog

  • Bay Watcher
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #336 on: March 12, 2023, 01:47:21 pm »

Interesting thought, although at the moment I mostly hunted by Microsoft word paperclip expressing love for me, asking me to leave my wife

Any idea how processing intensive natural language processing is? it can already solve the problem of webcrawler keyword gaming, but I have no idea how efficient it is at handling the vast amount of information out there.

You seem to have good memory, atm I can't seem to recall a single thing about asimov books not a book title, not character name, not even plot point. Maybe someone can come up with an AI that can help find that string of memory you can latch on to unravel stuff.


Does anyone know the extent and circumstances of ChatGPT "hallucinations" mentioned in the debate, because if it can just make up facts then it undermines my assessment of its usefulness even for simple facts.
Logged

Starver

  • Bay Watcher
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #337 on: March 12, 2023, 03:44:31 pm »

...initially unrelated to the (adopted) subject at hand, but:
You seem to have good memory, atm I can't seem to recall a single thing about asimov books not a book title, not character name, not even plot point. Maybe someone can come up with an AI that can help find that string of memory you can latch on to unravel stuff.
It is said that as well as intelligence (which is by far not unique to humans), and the basic shared intelligence of society (which apes and corvids and cetaceans and others demonstrably have), through the development of writing (and now other media) we have extelligence. A store of knowledge that does not necessarily rely even on being in the presence of others in the same time and place.

(So can we say that a "prompter" AI that whispers unsought facts can be yet another layer?,Or indeed does the collective virtual presence available through the intenet already do this, but maybe of a different degree of usefulness, given how wrong some such sources can be. As can all other layers of intelligence, but now somewhat lessening the consequences in the event of deliberate or accidental misinformation[1], the come-back being pretty much non-existent (or the point). And once we start to rely on the little AI daemon whispering in our ears, telling us what to believe... Well, maybe that's when we start to lose intelligence and they can be truly said to be the thinkers on the planet.)


As for my knowledge of Asimov, call me a past afficionado. It helped that Asimov was near the start of the SF shelves of the local library, when I was both young and rigidly methodical. Plenty of Asimov, Bradbury and Clarke. Not quite as much Verne, Wyndham and Zelazny, in comparison, except that Wells somehow got fairly well read despite this. ;)

(To clue you in, the "very near future" Robot-era books tends to be short stories that often feature Susan Calvin, robot 'psychiatrist' (very serious, more empathic with robots than people; except for telepathic robots!), or sometimes Powell and Donova (engineers, more comic relief to them). The stories in this area deal a lot with "This robot isn't doing what it ought to be doing. Why?", and the answer often being some inadvertent (or sometimes deliberately invoked, but still not intentional) edge-case where the robot is doing exactly what it should do, given the Three Laws being in the (mostly) proper order. The solutions tend to vary from "deal with it" to "do something about it, or you'll be sorry", with the occasional self-solving issue that still needed an explanation. The Zeroth Law comes into play as of the slightly later Spacer-era section (Novels) of the in-universe chronology and goes all the way to the far-future Foundation setting as a major background plot-point, it turns out, even though there are are no obvious robots...)


[1] Showing the young hunter the wrong way to hunt a mammoth isn't going to help the elder hunter much when all he's reliant upon the next generation or two when he isn't there to do the right things. A book that's not fit for reading can become firelighters or be hung up in the privvy and let it be useful in other ways (if the material paper is worth saving at all).
Logged

EuchreJack

  • Bay Watcher
  • Lord of Norderland - Lv 20 SKOOKUM ROC
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #338 on: March 12, 2023, 04:19:00 pm »

Let's be honest, the AIs being created have no ethical subroutines.  They have "niceness protocols" to avoid shocking the public, but no morality is being hardwired into them. Nor should it.

King Zultan

  • Bay Watcher
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #339 on: March 13, 2023, 03:33:59 am »

But how can they evolve into real boys if they don't have morals or ethics?
Logged
The Lawyer opens a briefcase. It's full of lemons, the justice fruit only lawyers may touch.
Make sure not to step on any errant blood stains before we find our LIFE EXTINGUSHER.
but anyway, if you'll excuse me, I need to commit sebbaku.
Quote from: Leodanny
Can I have the sword when you’re done?

Starver

  • Bay Watcher
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #340 on: March 13, 2023, 04:42:04 am »

*sound of cricket chirping*
Logged

jipehog

  • Bay Watcher
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #341 on: March 13, 2023, 05:13:10 am »

Blake Lemoine tears? (the Google engineer that thinks its LaMDA AI has come to life and was fired)

Btw hallucinating AI might be good at whose line is it anyway style hoedown.

@Starver, I was thinking more like a personal assistant, all your data stored in one places plus extra functions to help memory recall and stave of Alzheimer.
Logged

scriver

  • Bay Watcher
  • City streets ain't got much pity
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #342 on: March 13, 2023, 06:02:01 am »

The less we use our memory the weaker it becomes
Logged
Love, scriver~

anewaname

  • Bay Watcher
  • The mattock... My choice for problem solving.
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #343 on: March 13, 2023, 09:50:33 am »

An AI could be excellent in building/re-building memory pathways, in the same way a parent/mentor helps build memory pathways through communications/interactions, simply because the AI would being willing to go through the steps over and over of pushing the human to observe, consider, and plan actions. In that way, an "AI whisperer" could be very useful.

But how would one keep that AI's "platform" away from the control of other people who want to bind the AI's users to a position below them in the human hierarchy? I mean that a burden will be placed on the users of that AI, in the form of a tax or a service fee, in some form of currency that the users will need to pay (time, money, labor, blood, etc), and that burden will increase over time until the bulk of the AI's users have a shortage in that currency and one small group of jerks have a surplus.
Logged
Quote from: dragdeler
There is something to be said about, if the stakes are as high, maybe reconsider your certitudes. One has to be aggressively allistic to feel entitled to be able to trust. But it won't happen to me, my bit doesn't count etc etc... Just saying, after my recent experiences I couldn't trust the public if I wanted to. People got their risk assessment neurons rotten and replaced with game theory. Folks walk around like fat turkeys taunting the world to slaughter them.

King Zultan

  • Bay Watcher
    • View Profile
Re: Reactions to Chinese White Balloon of AI Awareness
« Reply #344 on: March 14, 2023, 02:11:08 am »

Oh no the AI will make us its slaves with its AI powers, and I for one do not accept our future robot overlords.
Logged
The Lawyer opens a briefcase. It's full of lemons, the justice fruit only lawyers may touch.
Make sure not to step on any errant blood stains before we find our LIFE EXTINGUSHER.
but anyway, if you'll excuse me, I need to commit sebbaku.
Quote from: Leodanny
Can I have the sword when you’re done?
Pages: 1 ... 21 22 [23] 24 25 ... 29