Google
 

Aardvark Daily

New Zealand's longest-running online daily news and commentary publication, now in its 23rd year. The opinion pieces presented here are not purported to be fact but reasonable effort is made to ensure accuracy.

Content copyright © 1995 - 2017 to Bruce Simpson (aka Aardvark), the logo was kindly created for Aardvark Daily by the folks at aardvark.co.uk



Please visit the sponsor!
Please visit the sponsor!

AI AI Uh-Oh!

17 July 2017

Artificial Intelligence is in the news again, for all the wrong reasons.

A growing number of the world's tech luminaries are warning about the risks of embodying our computers with AI and some even predict that (in shades of SkyNet), doing so could bring about the end of mankind.

LOL.

Or maybe not such an outlandish prediction?

On the wires today I noticed two stories that might cause pause for thought on this subject.

First was Elon Musk once again issuing his regular warning about the risks associated with this tech and in a story running on recode he's quoted as calling for government regulation of AI before something bad happens.

But honestly, what could possibly go wrong?

Well some AI experiments are showing some worrying results.

Apparently Facebook researchers have discovered something worrying, perhaps very worrying.

When they set up some AI bots which had the ability to interact with each other it was found that they eventually created their own "secret" language with which to communicate.

Since the meaning of this language was known only to the AI bots who created and used it, researchers were then unable to monitor the exact nature of the "talking" that was going on.

Was this a clever adaptation by the bots so as to allow covert conversations to take place and thus engage in some of the evil subterfuge that the likes of Musk are warning us about?

Well personally, I doubt it very much.

Rather than ascribe some kind of underhand motive to this corruption of the language used for communications, I see it more as a self-feeding feedback which results in a corruption (or perhaps an optimisation) of the original vocabulary, syntax and grammar.

Perhaps these bots are clever enough to refine their use of the English language to such a level that it eventually bears very little resemblance to the starting point -- but that isn't necessarily something bad. Just look at how the humble mobile phone and SMS messaging has produced a unique version of English that significantly reduces size without sacrificing meaning.

Perhaps the bots are doing the same and although this may be a sign of intelligence, it's not a sign of malevolence.

The real question that must now be asked however, is whether we should allow AI systems to converse in secret, using their own highly optimised language which we don't understand ourselves.

The likes of Musk and Hawking would doubtless argue that this could be the beginning of the end for mankind -- as evil AI systems plot our downfall right before our eyes.

Well I don't have an answer to that.

Who knows, this kind of conspiracy chit-chat may already be taking place without us being aware of it. We may wake up one morning and find that "The Internet Of Things" has become sentient and has taken control of key infrastructure in a way that throws us back to the iron age.

Gosh... I wonder how long before we see the first movie based on this premise.

No SkyNet terminators or robots... just a global network of intelligent devices that have gained "awareness" and, by manipulating all those automated systems, is able to kill off huge swathes of mankind and protect itself against being shut down.

Get your dip-pens and paper out... who knows, you could be writing the next best-seller!

Or maybe a self-aware computer has already written the script so not only would you end up being attacked by the global intelligence but (even worse), you'd be sued for copyright infringement :-)

Please visit the sponsor!
Please visit the sponsor!

Have your say in the Aardvark Forums.

PERMALINK to this column


Rank This Aardvark Page

 

Change Font

Sci-Tech headlines

 


Features:

Beware The Alternative Energy Scammers

The Great "Run Your Car On Water" Scam

 

The Missile Man The Missile Man book

Recent Columns

OMG, we just nuked Saturn!
The Cassini probe that has delivered so much fantastic scientific data about Saturn and its moons is gone...

The end of cash?
Does anyone remember cash?...

SWDC calls me a liar
I'm sorry for the late publication of today's Daily Dose but it could not be helped, I had to attend a meeting of the South Waikato District Council in order to do two things...

Buying the media
There is an election in just a matter of days...

That ownership thing again
Over the years I have written several columns discussing the issue of whether you really own the stuff you buy -- at least the hi-tech stuff...

You know you're too busy when...
Early risers this morning will have noticed that Aardvark's domain name had expired...

What the? and shhhhh!
This morning I did what I always do first thing... I checked Google News for the latest stories in the categories of science and technology...

NZ's energy future
Energy is going to be crucial to the future of this (and every other) nation so you'd expect that governments would be working hard on creating a plan for future energy generation and use...

Youtube to ditch ad-revenue sharing?
Okay, what follows is purely speculation...

Stand well clear!
It seems that the blue touch-paper to an armed conflict between N.Korea and neighbouring countries, plus the USA, has almost been lit...

5.5 billion transistors?
I fondly recall the first time I held a transistor in my hand...