AI just passed the Self-Awareness test

[bill flibby]

Here's the link.

What do you guys think about this? I've only done a little reading about artificial intelligence, and this article didn't really include a lot of details (as far as coding and algorithms goes) but I think this is pretty neat. I don't know if I would call this self-awareness, but the possibility seems pretty groundbreaking, if true.

 

[wakachamo]

Sensationalist title; this is far from 'self-awareness' — this is just a demonstration of simple reasoning techniques in an area known as First-order logic. You can be given a set of statements (or rules), relationships can be made between those statements, and therefore you arrive at new conclusions.

There is at least one robot who speaks —> I'm able to speak —> Nobody else spoke —> I'm the only entity able to speak.

The ability to solve such a problem is by no means an indication of 'self-awareness'. Impressive efforts in general, sure, but this isn't close to a machine resembling consciousness at all.

 

[esperance]

Haha, this is hardly what I would call self-awareness. It's simply a matter of sound recognition and basic reasoning, as wakachamo said.

Also, just look at the comments for the article. I nearly died.

Note this ludicrous test is sufficient to fool the average journalist.

 

[bill flibby]

Thanks for raining on my parade, guys.

For real, though, I wouldn't really call it a feat of self-awareness, either. I find it neat that we're at the technological point where we can create programs capable of working through first order logic. I mean, sets of rules and procedures are what programming languages consist of, but this is the first time I'm hearing of a program that uses an additional system of logic to perform its functions - in this case, the "self-awareness test" of recognizing its own noises, despite being "told" it can't make noises.

and I really want to see AI before I die, so any time I see something like this in the news I get excited hoping that progress is being made

 

[Veridio]

Yeah, it's really not that impressive sadly. I'm hoping we see some improvements within 10 years.

 

[Leviathan]

I'll start to worry when an AI achieves Skynet-like capabilities, thanks.

not bad for famous last words huh?

 

[Legobricks]

I'll start to worry when an AI achieves Skynet-like capabilities, thanks.

not bad for famous last words huh?

How would that be worrying? I gather you're joking but people do seem to profess this as a deeply held fear. At the end of the day, your human neighbour/friend/lover etc. is much more likely to snap and come for your blood. A mechanical mind may be devoid of conscience, but it is at the same time devoid of the processes that make it an emotionally-ruled, irrational mess, and violent tendencies don't pay the rational.

 

[Touched]

How would that be worrying? I gather you're joking but people do seem to profess this as a deeply held fear. At the end of the day, your human neighbour/friend/lover etc. is much more likely to snap and come for your blood. A mechanical mind may be devoid of conscience, but it is at the same time devoid of the processes that make it an emotionally-ruled, irrational mess, and violent tendencies don't pay the rational.

Emotions aren't what make a general intelligence dangerous.

 

[Flubbles]

This is actually very cool.