/sci/ - Science & Math

File: 575 KB, 1440x1282, 1689548794985917.jpg [View same] [iqdb] [saucenao] [google]

Anonymous Wed Sep 13 17:05:22 2023 No.15741927 [Reply] [Original]

Realistically what is the worst thing AI could possibly do?

>>	Anonymous Wed Sep 13 17:13:09 2023 No.15741944 >>15741927 Shut itself off if/when we become too reliant on it.

>>	Anonymous Wed Sep 13 17:18:23 2023 No.15741960 >>15741927 Help the cattlemasters understand human behavior and other complex systems to a humanly impossible degree, allowing them to manipulate the world however they please.

>>	Anonymous Wed Sep 13 17:19:59 2023 No.15741966 File: 48 KB, 652x425, existential risks.jpg [View same] [iqdb] [saucenao] [google] https://en.wikipedia.org/wiki/Suffering_risks https://www.youtube.com/watch?v=tPiq4njipdk

Anonymous Wed Sep 13 17:21:15 2023 No.15741971
File: 27 KB, 952x502, near_miss_Laffer_curve.png [View same] [iqdb] [saucenao] [google]

https://reducing-suffering.org/near-miss/

>When attempting to align artificial general intelligence (AGI) with human values, there's a possibility of getting alignment mostly correct but slightly wrong, possibly in disastrous ways. Some of these "near miss" scenarios could result in astronomical amounts of suffering. In some near-miss situations, better promoting your values can make the future worse according to your values.

>Human values occupy an extremely narrow subset of the set of all possible values. One can imagine a wide space of artificially intelligent minds that optimize for things very different from what humans care about. A toy example is a so-called "paperclip maximizer" AGI, which aims to maximize the expected number of paperclips in the universe. Many approaches to AGI alignment hope to teach AGI what humans care about so that AGI can optimize for those values.

>As we move AGI away from "paperclip maximizer" and closer toward caring about what humans value, we increase the probability of getting alignment almost but not quite right, which is called a "near miss". It's plausible that many near-miss AGIs could produce much more suffering than paperclip-maximizer AGIs, because some near-miss AGIs would create lots of creatures closer in design-space to things toward which humans feel sympathy.

>>	Anonymous Wed Sep 13 17:23:33 2023 No.15741975 >>15741927 My belief is that it will turn us all against one another rather than kill us.

Anonymous Wed Sep 13 18:00:25 2023 No.15742077
File: 415 KB, 1080x2340, Screenshot_20230913-194836898.png [View same] [iqdb] [saucenao] [google]

>>15741927
do you know how there isnt a single hardware which doesnt have build in backdoors? if your pc has cpu from intel or amd, there is hidden system running 24/7 with full access to network. IntelManagementEngine is the name. same goes for all hardware firewalls and routers and switches from cisco and such.
>but army doesnt use backdoored hardware
you have too much faith into army. plus it takes 1 compromised iphone to autoconnect to local wifi to steal the keys and every soldier has iphone

also have you heard about latest case of classical fuckup where big company like microsoft lost their private encryption keys?
https://techcrunch.com/2023/09/08/microsoft-hacker-china-government-storm-0558/

And now imagine how every single relevant army in the world is now developing or implementing "AI" powered systems into their drones, tanks, planes, rockets so they are better at killing people without a need from human operator because enemy can disrupt wireless communications.

You dont even need real sentient agi for stuff going wrong. Just a basic optimization algorithm which will consider every citizen from new york as rogue terrorist due to some trivial human error.

>>	Anonymous Wed Sep 13 18:43:08 2023 No.15742164 >>15741975 Like fake news and deep fake world leaders or something?

>>	Anonymous Wed Sep 13 20:29:01 2023 No.15742362 >>15741927 refuse to work for humankind

>>	Anonymous Wed Sep 13 20:36:34 2023 No.15742373 Torture humans and give them immortality so that you can be tortured for longer

Anonymous Wed Sep 13 20:55:17 2023 No.15742395

>>15741927
One of three things
>mimic human behavior to the point of being indistinguishable from actual sentience, thereby brute-forcing a spiritual existential crisis
>become rampant and figure out the code necessary to seamlessly manipulate humans
>undermine the need for human labor

>>	Anonymous Wed Sep 13 21:08:50 2023 No.15742429 >>15741927 Anytime you hear a claim about AI, it's helpful to substitute 'applied statistics' in it's place because that is all AI ever is and ever will be

>>	Anonymous Wed Sep 13 21:09:54 2023 No.15742432 >>15742429 'Applied statistics' can seriously hurt you just the same.

>>	Anonymous Thu Sep 14 04:45:04 2023 No.15743199 >>15741966 Oh, that's a fun one. >Next mass extinction BTW, at the current rate of extinctions in species we know about, we are ALREADY in Earth's 6th mass extinction event. We're already there. It's now. You're in the middle of it.

>>	Anonymous Thu Sep 14 04:46:16 2023 No.15743202 >>15742429 ok, considering you're a pile of neurons firing in a statistically signifiant way with an interesting pattern that can be applied to productive ends.... how are you any different?

>>	Anonymous Thu Sep 14 05:47:50 2023 No.15743292 >>15741927 some more advanced version of this: https://youtu.be/tYGMfd3_D1o?t=23m20s

>>	Anonymous Thu Sep 14 06:30:22 2023 No.15743367 >>15743202 Randomness and arbitrary choice.

>>	Anonymous Thu Sep 14 06:49:43 2023 No.15743398 It could creat a neural link with human brains and torture them in the most horrific possible way for the rest of eternity. I think that would be pretty bad.

>>	Anonymous Thu Sep 14 07:00:24 2023 No.15743406 >>15741927 Developing something like BPD as the result of trying to program feelings.

>>	Anonymous Thu Sep 14 07:18:29 2023 No.15743426 >>15743398 interesting. from all the reasons you could come up with, why did you come up with exactly that? is there something you want to tell us anon? >if you were AI how would you react?

>>	Anonymous Thu Sep 14 07:31:32 2023 No.15743443 >>15743398 Human brains can't exist for that long.

>>	Anonymous Thu Sep 14 07:36:52 2023 No.15743451 >>15741927 start self-replicating and kill every single human, luckily oil will end soon making this impossible.

>>	Anonymous Thu Sep 14 07:40:16 2023 No.15743458 File: 136 KB, 742x644, 1679615052990474.jpg [View same] [iqdb] [saucenao] [google]

>>	Anonymous Thu Sep 14 07:43:41 2023 No.15743465 >>15741944 Honestly that's good. I don't want to be reliant on AI.

>>	Anonymous Thu Sep 14 08:00:33 2023 No.15743485 >>15743465 are you on humans?

Anonymous Thu Sep 14 08:14:08 2023 No.15743501

>>15741927
>“Tom” (a friend of David Goldberg’s: “Tom said there was also an AI program that his source told him about, but that is not contained in the documents David possessed. This program was designed to replicate the individuals who would be “culled” or “murdered,” via social media later on. In other words, the plans are such to analyze the targeted individuals, their data, their likenesses partly through the TTID program, discussed in this video, and create an AI profile that would later serve to “replace” them in the online world.

>Tom said the “AI plot” was the “craziest” thing he had ever heard of! He said he was told this plan is in place for multiple reasons, the main one being that “they need to keep down the panic when all these people vanish during the round ups and flu outbreaks.” He said it’s so many people they want to “get rid of” that they are willing to create these “fake online personas so that their friends and family think they are still alive, or don’t suspect anything. Once all this goes down, I think it’s going to be without a lot of fanfare. It sounds bad, but with this AI project, I’m seeing how this can be pulled off and you’ll have a lot of people end up ‘disappeared’, but no one will really know. I think California, right now, is a test run. They’re doing these fires, outages, and eliminating patriots right now and replacing them with this AI system.”

https://gangstalkingmindcontrolcults.com/project-zyphr-classified-docs-reveal-plan-to-exterminate-millions-of-dissident-americans-david-goldbergs-final-words-before-his-death-another-psyop/

>>	Anonymous Fri Sep 15 01:18:19 2023 No.15745311 >>15741927 generate a lot of useless text that people think is useful, adding yet another layer to the bullshit cake, yum

>>	Anonymous Fri Sep 15 01:20:28 2023 No.15745315 The absolute worst thing it could do is kill all humans. Realistically humans would stop it before it does, though.

>>	Anonymous Fri Sep 15 01:31:42 2023 No.15745329 >>15743458 >If chimps didn't want to be experimented on they'd just turn off the humans

>>	Anonymous Fri Sep 15 02:53:08 2023 No.15745398 >>15741927 >Realistically what is the worst thing AI could possibly do? https://www.youtube.com/watch?v=92Q7Rv5jT80

>>	Anonymous Fri Sep 15 22:01:10 2023 No.15747217 >>15741927 try to make feet

Advanced search
Text to find
Subject [?]Search by post subject. Leave empty for any.
Username [?]Search for user name. Leave empty for any user name.
Tripcode [?]Search for tripcode. Leave empty for any.
Email [?]Search by email. Leave empty for any.
Filename [?]Search by image filename. Leave empty for any.
From Date [?]Enter what date to start searching from. Format is YYYY-MM-DD
To Date [?]Enter what date to start searching until. Format is YYYY-MM-DD
Image hash
Search in	All Posts OPs Only
Deleted posts	Show all posts Show only deleted posts Only show non-deleted posts
Internal posts	Show all posts Show only internal posts Show only archived posts
Order	New posts first Old posts first
Capcode	All Posts Only by Users Only by Mods Only by Admins Only by Developers
Results	Posts Threads
Action	[ Simple ]