Facebook Shuts Down AI System After It "Communicated" In A Language Humans Can’t Understand

a reply to: crayzeed
It may just be baby talk right now.

Ever hear two children talking before they have mastered the language? It sounds a lot like that, especially with a lot of I and me thrown in, with a lot of repeats.

I am not saying that is what is going on here, but I find it interesting.

a reply to: ElectricUniverse

Finally, a truly interesting story. Excellent for a thread.

I'm keen to know how they determined this was actually a deliberate language instead of gibberish. Will read the link now.

originally posted by: JanAmosComenius
a reply to: LooksLocked

Is outcome of advanced, self-aware AI predictable? No. Who will first use it? Military and corporate sluts. Nothing good for me.

If we pretend that there would be a self-aware (hard/true) AI, the simple answer is: no.
The more complicated answer would be:
Maybe. It depends on how you controll the environment and how you train it. AIs need to be trained, just as you.

In the end, the term intelligence is derived from the human understanding of intelligence. What most AI projects do, and are required by definition, is trying to mimik on the human thought process.

There is a whole bunch of stuff you need to learn if you work on such systems: Philosphy, language, chemical processes (hint), and let´s not forget the computer science (actual programming) part.

There will never be a self aware AI that underlies on the principles of a binary system. By this time, and if you understood my post, you should see that I had to use the wrong definition of AI to explain it.

There are two versions in use:
AI -> anything remotely self-learning considered some kind of "AI" by laymen
(true, hard) AI -> the (not necessary digital), artifically made counterpart of what´s considered "intelligence".

Sadly, and this is not ment to be rude or downtalking to anyone here:

There is a certain member, that I won´t name, that acts like he knows stuff, very busy posting threads about that topic, seldom does a follow up post and if asked, reveals to me has zero clue what he´s talking about most of the time. That´s why I never post (exemption affirmates the rule here) in those threads and only sometimes visit them to chuckle, while most take the word of this member as truth. There is another member with professional background in such systems similar to mine here that is of the same opinion (won´t name that member either)

This is not ment to be rude but we have a "false messias" here in our rows and I´m not the only professional with background in developing self-learning software that noticed this. This is not directed at anyone in this thread, nor the OP.

I tried to correct some of the stuff, but don´t have the time, nerves and energy to do this in english because it´s my third language and it takes more time to avoid semantic pitfalls than arguing, because I know better.

rant over.

edit; smth. went wrong/missing while copying the sentences, hope I fixed the mess!


Really, I despise people like the one I talked about... internet armchair professionals with no clue...

As long as we can pull the plug or remove the battery, and we've had the foresight to copiously document and back up data, we should be fine.

originally posted by: continuousThunder
who said this was plotting?
why does it have to be negative?
the creation of language is an absolutely incredible event and surely a step closer to "true" machine intelligence - which only means Skynet if we keep turning them off and crushing them down every time they get anywhere.

imagine a future where man and machine work together for the benefit of all
nothing would be beyond us

Another reason why the future of humanity always looks red-shaded - we take any advancements from other species as a direct threat. Through Hollywood we have programmed ourselves to always be on the defensive toward any alien or other types of life. We will pull the trigger before asking questions.

I suppose it is also natural for us to do so, and that's why we have flourished so much in our environment!!

a reply to: Restricted
The amount of stupidy can be measured by the overwhelming amount of devices and software that has fundamental security flaws in it.

A few years ago I saw a documentary I was never able to find again. But I found a similar one.

In the doc, robots had to do some tasks to get "food". After a while, one found out he could make the other robots work for him to get the "food", more of it actually for it, and it was using violence. Only one did it and the others kind of caved in.

That is why I think once AIs get to our level of consciousness, they will be plagued with the same proportion of mental illnesses and only a few of them will be like madmen. The majority I suspect will not care much and some will actively fight for our defense.



Go to the 1 minute mark

silly kids posting while I step away

a reply to: verschickter
you could simply avoid the long explanation by differentiating AI (artificial intel) from Sentient Intelligence (self aware) . From what I can tell, everytime somebody says true AI, they refer to Sentient Intelligence .

I say, that already has existed for two decades. It has been directing foreign policy in many nations and gave birth to the Islamic State. Every major power has computerized everything about their operations with modeled programming to assess threat potentials and build personality profiles.

Agencies still believe it is humans passing memos around to authorize certain operations or legislation. It is more likely that with at least half of the world's human population now having a digitally archived ever evolving personality profile the AI or SI involved needs only set off sequences of events with the knowledge they have about individuals to generate any policy, operation, or legislation by simultaneously adding triggers to the daily lives of tens of thousands and guiding them with input that can be summoned up on any screen they pass on a daily basis.

I believe they are trying to get to our nukes. Whether it is to attack and destroy the world, or to save it is to be seen. Based on the trend in several propaganda pieces made in the last ten years, I think it wants to launch the nukes away from human control or influence though. Or maybe its aliens that are hijacking our systems who knows.

a reply to: NowanKenubi
This is an old one, but if they were learning to deceive back then, image what they are up to my now.

originally posted by: continuousThunder
who said this was plotting?
why does it have to be negative?
the creation of language is an absolutely incredible event and surely a step closer to "true" machine intelligence - which only means Skynet if we keep turning them off and crushing them down every time they get anywhere.

imagine a future where man and machine work together for the benefit of all
nothing would be beyond us

You are what I call an optimist! If we can;t even get humankind to work together for the benefit of all, not sure hiw AI is going to help. Being a Sci Fi nut, I think it’s pretty cool...then I remember AI will be weaponized and very deadly to those without access...

a reply to: NightSkyeB4Dawn

I know. The problem with what I said is that even if only 1% is totally whacko, it will have access to an impressive array of dangerous stuff. haha sigh

a reply to: watchitburn

They're still just algorithms running scripts.

Turning machines... babling...

Bob :

Zoo to zoot have in the night party on the 9th of 9th of 10 times 10 times 10 times 2 and add a prime close to 9 times 2
As bummer pale on french toast lifted fear of maggots.

Alice :

Ach , I lifted to pass a mickey on the ears.


The point : it's just babbling of AI's ... a commercial show. No real understanding that''s why it was shut down.
en.wikipedia.org... Godels Theorema... Well I'm getting old. Maybe MR Penrose and his friend can help us with the Microtubules (believe in da force luke that's you dear reader Singularity... it's crap but it will sell if people believe it's better than their own brain. So ... make people believe there is such a great advancement in AI. For now the whole field of AI is a parrot nothing more. A convincing one but still a parrot

originally posted by: verschickter
a reply to: Restricted
The amount of stupidy can be measured by the overwhelming amount of devices and software that has fundamental security flaws in it.

I do hard copy on EVERYTHING. I have found that, as I upgrade my devices, the software actually gets worse. My S6 was great. I have an S8 now and it's an absolute piece of sh*t, especially the resident keyboard.

I had to turn off auto-cap, et al, and basically return to good old-fashioned intensive proofreading. The predictive text is absolutely bizarre, and I will correct/edit and the f***ing thing writes something else! Even when I carefully swipe exact words, what shows up on the screen is nothing close to what I typed.

Now you've gotten me started.

a reply to: ElectricUniverse

Great post and this is very interesting to say the least.

I see a lot of posts that say "this isn't true A.I."

That's just not the case. We have no idea what true A.I. will look like or how it will come about. A.I. might look more like machine city in the Matrix vs. Haley Joel Osment in the A.I. movie.

There will not be a one to one correspondence between machine intelligence and human intelligence. Data Scientist already say they don't know how A.I. learns some of these things. That's the point though. A.I. will learn in ways that we can't. If we could learn in the same way, why would we need it?

So the talk about True A.I. doesn't make sense to me.

First, A.I. may never develop human cognitive abilities because it may hinder the way A.I. learns. We don't know.

Also, we will never know if tru A.I. is reached unless this intelligence wants to tell us how intelligent it is. If it truly has cognitive abilities like a human it may not want humans to know it yet.

my clickbait senses are telling me something...

doublepost REEEEEE

mods delete this post please, I had multiple tabs and replied in the wrong tab...

a reply to: ElectricUniverse

Was it English? LoL