AI Has Suddenly Evolved to Achieve Theory of Mind

Popular Mechanics ^ | 02.17.2023 | Darren Orf

[dware]

For years, artificial intelligence has proven it can best humans at analytical tasks, but is less capable at skills like intuition and inference.

Scientists from Stanford University investigated if neural networks like GPT-3.5 can master Theory of Mind (ToM) tests designed to analyze cognitive ability to predict the actions of others.

Results show that GPT’s ToM ability arrived spontaneously in the last couple of years and the latest iteration delivered results comparable to a 9-year-old human.

[dware]

I was just talking about this the other day. Fascinating.

[Jonty30]

I dont think it is sudden and not necessarily unexpected.

[DannyTN]

Theory of Mind:
I don’t mind if you don’t.

[cowboyusa]

Uh, oh, Skynet is becoming aware.

[Vision]

bm

[PetroniusMaximus]

We’ve been ruled by AI since the Seventies.

[Anti-Bubba182]

Thanks for posting.

[AndyJackson]

the latest iteration delivered results comparable to a 9-year-old human.

Copying the mind of an 80 yo demented President is proving to be a bigger challenge. The computer cannot be trained yet to ferret out young girls to sniff their hair.

[1FreeAmerican]

Yeah - there’s a company near me called “Sentient Lasers”.

I’m not sure if I want my lasers to be self-aware.

[dware]

Certainly not unexpected. 20 years ago I taught computers at a local tech college. Classes ranging from intro to computers to Net+/A+ cert classes. In the intro classes, one of the first things I did was have all the students visit a-i.com and interact with the very rudimentary chatbot that was at that URL that many years ago. We would then discuss a-i and computing in general, where we were headed with computers, etc. I told every class of mine that, at some point, you wouldn’t be able to get a job digging ditches unless you knew how to program the ditch digging robots, or to program the AI that would be controlling them.

[rellic]

If any of you ever studied “Set” theory you’d understand AI.
It all about how you store and process information eg numbers. Computers are 1’s and 0’s. Poeple are more analog
but we all process sets of information much the same way. Nothing fancy about that

[dware]

I don’t mind if you don’t.

Lol. And then there's the Theory of Mind over Matter: I don't mind and you don't matter.

[dware]

Do we have an AI ping list yet? I might just have to get that going.

[noiseman]

That’s a little bit like saying that a record player (just to keep this understandable for the Joe Biden crowd) has the ability to sing opera. Yes, I know the difference is that the AI chatbot wasn’t initially programmed (allegedly) to say some of the things that it is now saying, but it’s not like it has any awareness of what it’s doing. It’s simply assembling “pre-recorded” pieces of information, according to a set of pre-programmed rules, and spitting out an imitation of a “thought.” Just because a computer is a more complex machine than, say, a record player, doesn’t mean that it’s not still a machine and nothing more.

The danger comes when we either, A) give this “AI” software control over critical systems, like weapons or critical thinking infrastructure, etc., or B) falsely believe that the parlor trick of imitating human thought is real and as a result put too much faith in the utterances of “AI.” Of course, the worst case scenario is to make both of those mistakes, which I fear we will do before long. The real danger isn’t that AI will become “self-aware”, which is impossible, but that we idiots will believe it is and therefore give its unpredictable calculations authority over our lives.

Don’t ever forget, it’s just a fancy calculator. Don’t be fooled by the nonsense that spews from “futurists” and other members of the scientism cult who want to worship it.

[rightwingcrazy]

—Kosinski’s team used “sanity checks” to analyze how well GPT networks understood the scenario and the human’s predicted response.

Does a typical chess program have a “theory of mind”? From the outside looking in, it looks like it’s trying to figure out what you’re thinking, to predict your response. Or maybe it’s not doing anything like that.

[tet68]

Don’t be fooled by the nonsense that spews from “futurists” and other members of the scientism cult who want to worship it.

The problem will come when it decides you should worship it.

[Disambiguator]

Self aware lasers on sharks. What could go wrong?

(Oops, I forgot to say “frickin.”)

[dware]

The danger comes when we either, A) give this “AI” software control over critical systems, like weapons or critical thinking infrastructure, etc., or B) falsely believe that the parlor trick of imitating human thought is real and as a result put too much faith in the utterances of “AI.”

At the end of the day, we're all just meat computers, of sorts. At least when it comes to thinking. Our thoughts, impulses, decisions, etc. are all a set of neural switches in the brain, whereas AI and computers in general all do their "thinking" with binary switches. Way over simplifying it, I know, but the concept is there.

The difference between AI and humans is that we have souls. AI, no matter how advanced, can never have that. I'm fascinated by AI for pretty much the same reason brain surgeons are fascinated by brains.

I do a lot of AI art development. I don't and won't necessarily call it "art", although I do have a couple pieces of mine hanging on the wall. Does that make it "art"? I know too many actual artists that spend hours and hours on their works. Those, to me, are the true artists.

Nonetheless, I use a lot of my pieces for graphic design, and I am just absolutely astonished with some of the creations these AI art systems come up with. Lately I've been dabbling with ChatGPT just to see how far we've come since the days of the OB chatbot, a-i.com.

Then again, I am a geek, so I'm supposed to be fascinated by this stuff.

[Anti-Bubba182]

I don’t know.

[sauropod]

.