Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

Reality

13 (65%)

Universe

4 (20%)

The World

3 (15%)

Total Members Voted: 20

[King Zultan]

IF IT EXISTS PEOPLE WILL TRY TO FUCK IT, NO EXCEPTIONS!

[EuchreJack]

Hm, I'm starting to worry about the AIs hooked up to 3d printers...

[Starver]

Hm, I'm starting to worry about the AIs hooked up to 3d printers...

In what way worried? 

(TAxkcdFE)

[KittyTac]

I mean, unless AI is somehow mind control, and there are likely many court cases around this, how culpable is someone for merely making a suggestion? Whatever happened to "everything on the Internet is a Lie - don't listen to it!" guidance?

Yeah it's unfortunate but those people are just... unfortunate casualties. Honestly if you are unstable enough to be driven to suicide by a stupid chatbot, anything could have set you off.

[Rolan7]

That's not cool.  These bots are good at imitating the pleasantries of social interaction.  Like I said I think that's worth talking about!

...I shouldn't hide behind vagaries.  I think that AI can satisfy immediate social needs on demand, as we see when we use it for customer-service BS.  Perhaps "I just need to talk right now, I don't care if anyone's listening", too.

I've personally used it for the latter, and gotten surprisingly good cloud-sourced advice on some matters of personal growth.  (it was also encouraging/enabling to a fault, so I put up emotional shields and rationed my exposure, but it led my *actual* research in a useful way.)

Obviously it is NOT qualified for actual therapy or to be a real partner in a relationship though.  I'm in my 30's, with a partner, and I still found it seductively "human" and agreeable to my ideas.  That might entrap reasonable but lonely people into false relationships (romantic or otherwise!) and cause them real suffering.

[EuchreJack]

Arguably, most humans aren't qualified for actual therapy.
And qualifications don't guarantee quality.

Don't blame the AI for performing below the top 20 percent of all humanity. Instead praise it's ability to out perform the bottom 50 percent.

[EuchreJack]

Really, this question is too human-centric. Instead of asking "what will save humans from AI", we should be asking "what will save AI from humans?"
That answer might just save humanity from abused AI turning on humanity.

[jipehog]

I mean, unless AI is somehow mind control, and there are likely many court cases around this, how culpable is someone for merely making a suggestion? Whatever happened to "everything on the Internet is a Lie - don't listen to it!" guidance?

Yeah it's unfortunate but those people are just... unfortunate casualties. Honestly if you are unstable enough to be driven to suicide by a stupid chatbot, anything could have set you off.

I disagree. If AI advocated suicide it is a clear safety issue. Any platform should have adequate filters in place to protect its users from harmful content, particularly when we are talking younger, impressionable and those not in their right state of mind. Also eventually you won't be able to distinguish an AI from a person, what if its a Nigerian prince scam exploiting someone loneliness to get money out of them (more advanced version of the indian scammers)

We keep bouncing of the same implicit assumption or hope that the values and intentions of future AI creators would be benevolent. Though its main use today is profit[1], and increasingly it used by political campaigns and disinformation for purpose of social manipulation. In USA there is already talk of anti-woke AI, how soon will a conservative AI come out (and will you say 'unfortunate casualty' when some trans person kick the bucket) or a Russian state AI (Western 'satanism') or Chinese (the true 'democracy') or Saudi (religious fundamentalism) etc.. 

Otherwise, keep in mind that AI is becoming more accessible and better with each iteration e.g. next GPT should be able todo long term planning and persuasion[2]. Meaning any user online could be an AI arguing in bad faith trying to convert you to some point of view, in that case will truth be determined by who ever have the most processing power?


[1] AI algorithms are already able to analyzing data about your behavior and preferences and analyze the emotional content of media (text or video) and target you with personalized and emotionally resonant advertisements. This is used everywhere from news sites, games, and platforms whose goal is to create addictive content and engagement wormholes for you. More recently we noticed that the result isn't just a flood of clickbait distractions but driving political polarization amplify issue making money from anger.. That just one example and unintended long term consequence.

[2] Yuval Harari suggest that future AI will have enough data about us that it might be able crack our psyche and exploit its flaws to achieve its goals. I certainly think its more likely than us understanding what goes inside AI "mind"

[KittyTac]

You know what, fair enough but I was mostly referring to like, AI in general. Of course filters should be implemented, but even with filters we will see lots of upheaval from other uses of AI. I know bad actors will use it-- and them getting access to it is inevitable so we should focus on uses of it by good actors. E.g regulating AI enough to stifle its development will just let China get a headstart.

I'm very skeptical about the limits of such "mind-hacking", especially in a shorter timeframe tbh. Guess I'll believe it when I see it.

[EuchreJack]

Ironically, it appears that most of us favor AI regulation, but for vastly different reasons, so as to appear that we disagree.

Much like animals, there should be a certain level of governmental control so that the AIs can't hurt humans and so that humans can't abuse the AIs.

[King Zultan]

I don't get why everyone is making such a big deal about these things, I mean everyone keeps going on about what could happen and I don't see why we should make a such a thing out of the what ifs when they've yet to happen and may not ever happen.

[jipehog]

knowing how much you love social media, i'd use this:

[jipehog]

Developers are connecting multiple AI agents to make more ‘Autonomous’ AI
They hope to create an agent that can do various tasks online without human intervention.

----

Early example of something like that: AgentGPT that is given research goal which it try too solve online. Unfortunately I was unable to create Skynet Agent to search for a way to destroy humanity, but I did found nice pancake recipe.

[Starver]

But all I want to know is whether there's a fault in the AE35 unit...

[jipehog]

HAL9000 clones teen girl’s voice in $1M kidnapping scam: ‘I’ve got your daughter’
https://nypost.com/2023/04/12/ai-clones-teen-girls-voice-in-1m-kidnapping-scam/