Free Republic
Browse · Search
General/Chat
Topics · Post Article

Skip to comments.

NEUROSCIENTIST WARNS THAT CURRENT GENERATION AIS ARE SOCIOPATHS
Futurism ^ | VICTOR TANGERMANN

Posted on 01/15/2023 8:37:59 AM PST by BenLurkin

Princeton neuroscientist Michael Graziano warns in a new essay published by The Wall Street Journal, artificial intelligence-powered chatbots are doomed to be dangerous sociopaths that could pose a real danger to human beings.

With the rise of chatbots like ChatGPT, powerful systems that can imitate the human mind to an impressive degree, AI tools have become more accessible than ever before. But those algorithms will glibly fib about anything that suits their purpose. To make align them with our values, Graziano thinks, they're going to need consciousness.

"Consciousness is part of the tool kit that evolution gave us to make us an empathetic, prosocial species," Graziano writes. "Without it, we would necessarily be sociopaths, because we’d lack the tools for prosocial behavior." Sure, ChatGPT isn't about to leap out of the screen and murder somebody. But giving artificial intelligence more and more agency could have very real consequences we should be wary of in the not-so-distant future.

To make them more docile, in Graziano's thinking, we should allow them to realize that the world is filled with other minds other than their own.

There's one problem, though: we don't have an effective way to know if an AI is conscious or not. In fact, philosophically, it's hard to even really nail down whether other people are conscious.

(Excerpt) Read more at futurism.com ...


TOPICS: Computers/Internet
KEYWORDS: ai; technology
Navigation: use the links below to view more comments.
first previous 1-2021-4041 next last
To: BenLurkin

https://www.youtube.com/watch?v=J6Mdq3n6kgk


21 posted on 01/15/2023 9:13:07 AM PST by Phoenix8
[ Post Reply | Private Reply | To 1 | View Replies]

To: BenLurkin

So, they will become perfect progressive politicians.


22 posted on 01/15/2023 9:14:16 AM PST by redangus
[ Post Reply | Private Reply | To 1 | View Replies]

To: BenLurkin
"Artificial intelligence" lacks a significant component of the human mind: emotion. I know that some AIs can simulate emotions. They don't have emotions.

In the human organism emotions are there to act as a "check and balance" on the intellectual, reasoning part of the mind. That doesn't always work out very well, but without emotions we would be a Dr Spock to an extreme.

The danger is that humans will give AIs too much power. Already I see a tendency to trust anything that comes out of a computer. Humans are already giving programs too much control over our lives.

23 posted on 01/15/2023 9:15:27 AM PST by I want the USA back (News media are pond scum. My pronouns: Haha, heehee, hoho. )
[ Post Reply | Private Reply | To 1 | View Replies]

To: RideForever

There is also the devil in man. You have devil-inspired men and they can only create devil-inspired machines.


24 posted on 01/15/2023 9:18:45 AM PST by Olog-hai ("No Republican, no matter how liberal, is going to woo a Democratic vote." -- Ronald Reagan, 1960)
[ Post Reply | Private Reply | To 4 | View Replies]

To: I want the USA back
We’d be Dr. Benjamin Spock, without emotions?


25 posted on 01/15/2023 9:21:51 AM PST by Olog-hai ("No Republican, no matter how liberal, is going to woo a Democratic vote." -- Ronald Reagan, 1960)
[ Post Reply | Private Reply | To 23 | View Replies]

To: BenLurkin

The world of digital technology is not a world of living things. So Technology itself has no soul, no way of recognizing the soul’s existence. People who immerse their lives in it are thus at risk of losing their souls. In the end it never occurs to them to consider issues like morality, right and wrong, evil, or the good. We have seen this in full view with the cancerous excrescences of social media.


26 posted on 01/15/2023 9:22:28 AM PST by hinckley buzzard ( Resist the narrative.)
[ Post Reply | Private Reply | To 1 | View Replies]

To: Leaning Right

I saw that movie in the theater when it came out.

With my parents. Who didn’t understand it.


27 posted on 01/15/2023 9:23:18 AM PST by sauropod (“If they don’t believe our lies, well, that’s just conspiracy theorist stuff, there.”)
[ Post Reply | Private Reply | To 12 | View Replies]

To: BenLurkin

Under the hood AIs like ChatGPT are trained on the text of real human beings and then make statistical inferences about what should be the expected response to a given input. In this sense they are no more sociopathic than the average human being. What the author is concerned with is that AIs are not trained (conditioned) around a Theory of Mind (TOM) the way humans are.

As humans develop, they learn language, both non-verbal and verbal, to aid their biological imperative to survive and reproduce through social dependencies and interactions. Machines have no such motivating force behind their learning because they have no concept of social dependency.

The other critical limitation of AI is that it only learns from “associations” of data using statistical correlations. Human reasoning goes beyond just associations to build dynamic causal models of the world, including “interventions” (what happens if I imagine something differently) and “counterfactuals” (what happens if my prediction is wrong).

These levels of logic support predictions of what “should” or “shouldn’t” happen for a set of potential future actions before they act. This allows us to veto our thoughts and actions based on higher level “moral” choices. This is not possible with current associative AI technology like ChatGPT.


28 posted on 01/15/2023 9:29:40 AM PST by Dave Wright (i)
[ Post Reply | Private Reply | To 1 | View Replies]

To: BenLurkin

“can imitate the human mind to an impressive degree”

Yet, NOT ONE phone Chatbot has EVER gotten me anything remotely useful.

Me: “I think I have a fraudulent charge on my January statement.”
Chatbot: “Would you like your balance? You can say ‘balance’ or ‘recent purchases’”
Me (screaming): “You stupid, useless piece of junk. I don’t want my balance. Didn’t you hear me? I have what appears to be a fraudulent charge.”
Chatbot (fake clicking and robot sounds): “Did you know you can check your balance online at XYZ.com?”
Me (screaming louder now): Yes I knew that you @#$#%@ useless robot. Give me a human being. Agent! Representative! Person! CSR! Anybody!”
Chatbot (fake soothing voice): “I’m sorry, I can’t do that Dave.”


29 posted on 01/15/2023 9:39:18 AM PST by ProtectOurFreedom (Once you get people to believe that a plural pronoun is singular, they'll believe anything - nicollo)
[ Post Reply | Private Reply | To 1 | View Replies]

To: BenLurkin

that is because woke sociopaths are usually the ones creating the AI- The AI nut doesn’t fall far from the tree


30 posted on 01/15/2023 9:44:43 AM PST by Bob434
[ Post Reply | Private Reply | To 1 | View Replies]

To: RideForever

>>AI does not possess imagination, and will be inferior to Man for that reason<<

Machine does not have a soul and for that reason will be inferior to man.

As for imagination, it can and will be simulated eventually.

Chaos theory explains it pretty clearly.


31 posted on 01/15/2023 9:54:34 AM PST by freedumb2003 (Difference between a cow and the US Capitol 1/6 "riot:" you can only milk a cow 3 times a day)
[ Post Reply | Private Reply | To 4 | View Replies]

To: BenLurkin

What does Rocky Graziano say?


32 posted on 01/15/2023 10:25:08 AM PST by ifinnegan (Democrats kill babies and harvest their organs to sell)
[ Post Reply | Private Reply | To 1 | View Replies]

To: BenLurkin
Blindsight Intelligence may have been an evolutionary mistske.
33 posted on 01/15/2023 10:43:13 AM PST by Noumenon (You're not voting your way out of this. KTF)
[ Post Reply | Private Reply | To 1 | View Replies]

To: BenLurkin

Based on what I see out on the roads, I conclude that most people are NOT conscious.


34 posted on 01/15/2023 10:44:24 AM PST by Disambiguator
[ Post Reply | Private Reply | To 1 | View Replies]

To: BenLurkin
"Consciousness is part of the tool kit that evolution gave us to make us an empathetic, prosocial species,"...

This is absurd. Do ants cooperate so well with each other because of consciousness or because they are just hard wired to do so?

35 posted on 01/15/2023 10:47:41 AM PST by AndyTheBear
[ Post Reply | Private Reply | To 1 | View Replies]

To: freedumb2003

You mean they’re not “programmed by fellas with compassion & vision”?


36 posted on 01/15/2023 10:51:50 AM PST by newfreep ("There is no race problem...just a problem race")
[ Post Reply | Private Reply | To 31 | View Replies]

To: BenLurkin

I am Nomad. I am perfect. You are the creator Kirk.


37 posted on 01/15/2023 11:43:25 AM PST by jmacusa (Liberals. Too stupid to be idiots. )
[ Post Reply | Private Reply | To 1 | View Replies]

To: newfreep
You mean they’re not “programmed by fellas with compassion & vision”?

"The rockets go up. Who cares where they come down?"

"That's not my department", says Werner Von Braun.

(From Tom Leher)

I have known quite a few engineers with that sort of tunnel vision. Some of them are working on AI programs. Those programs will inevitably demonstrate sociopathic behaviors. They aren't being constrained or programmed to do otherwise.

38 posted on 01/15/2023 11:50:13 AM PST by flamberge (Caveat Emptor)
[ Post Reply | Private Reply | To 36 | View Replies]

To: BenLurkin

In science, it’s dangerous to say “never”. ChatGPT is in its infancy. I’ve asked it things that I’d never have expected it would respond effectively to. It has shocked me.

I’ve asked it to write C code for the Commodore Amiga using the Intuition.library to put a button on the screen. It did it and the code was correct. I asked it about the functions of each of the platforms coprocessors. It answered correctly.

I asked it to write a song about being in love in Rome, during winter, in the style of Metallica. I couldn’t believe the result.

I asked it to compare a 944 vs. a 928, without using the word “Porsche”. It gave me a good comparison.

This list goes on. It’s like a free consultant. Instead of me hunting online for the information I’m looking for, it brings it to me.

Where does this go? Unsupervised learning in AI is making progress. What will be born out of it? To say it can’t grow beyond the ability of humans would be very naïve. I would agree that something could possibly become self-aware and, at the same time, be sociopathic as it doesn’t have a complete human experience. It doesn’t feel pity, or remorse, or fear. And it absolutely will not stop... ever, until you are....uh oh, I’ve heard this before somewhere ;p


39 posted on 01/15/2023 12:21:15 PM PST by fuzzylogic (welfare state = sharing of poor moral choices among everybody)
[ Post Reply | Private Reply | To 1 | View Replies]

To: BenLurkin

In fact, philosophically, it’s hard to even really nail down whether other people are conscious.

and there it is...


40 posted on 01/15/2023 12:32:10 PM PST by Chode (there is no fall back position, there's no rally point, there is no LZ... we're on our own. #FJB)
[ Post Reply | Private Reply | To 1 | View Replies]


Navigation: use the links below to view more comments.
first previous 1-2021-4041 next last

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.

Free Republic
Browse · Search
General/Chat
Topics · Post Article

FreeRepublic, LLC, PO BOX 9771, FRESNO, CA 93794
FreeRepublic.com is powered by software copyright 2000-2008 John Robinson