Problem Solved: Unfriendly AI Monica Humanity HPlus

Jan 24, 2011 09:22 AM

article in hplusmagazine.com  on Dec 15 2010.

Monica Anderson

opening:  future chess playing computer can predict chess moves but gets destroyed

by fire sprinklers when some Grand Marnier gets spilled.

Nobody can predict that... world is too complex.

The purpose of intelligence is prediction.

Humans have used their minds to predict the behavior of sabertooth tigers, the right day in spring to plant the crops, and the behavior of opponents in games like chess and tennis.

Predictive capabilities actually predate chess by hundreds of millions of years.

An animal running on a rocky beach needs to predict where its legs will land and when to start activating the leg muscles to handle each impact.  Here, the prediction reaches only milliseconds into the future.  Often this is enough.

The ability to better predict what an opponent will do in a fight to the death

conveys the ability to parry a strike in time rather than milliseconds too late.  But better than that, the first rule of street fighting is "Be Somewhere Else".  Predicting that a fight is imminent is more useful than predicting the first blow.

Predation requires intelligence, but more superior intelligences develop safer techniques like scavenging, ambushing, pack hunting, weapons, traps, and Trojan Horses.

An evolutionary arms race has ratcheted intelligence levels ever since the first brain-like nerve clusters appeared.

Longer term predictions, predictions that are more often correct, and predictions that are more specific about the outcome all provide evolutionary advantages for superior predictors.  As our ancestors started living in larger tribes, the ability to predict the behaviors of potential mates and rivals made a big difference for who got to breed.

Intelligence-based mate selection was one key factor in the explosive increase of intelligence levels in our species.

The Libet Delay ... she claims its a half a second...  no , no , no wrong... but idea is right.

Predictive capability selects the winners among frogs darting their tongue against flies and among flies avoiding tongues.

The world is TOO COMPLEX for any intellig systems (let along logic-based system) to make longe range predictions.

But predict we must.  Planning is just a prediction that includes our own actions.

We quickly and subconsciously hypothesize a few alternative actions, predict some of their possible outcomes, and then select the action that we predict will bring about the most desirable outcome.  This is an error-prone process, but since it's the best we can do, this is what we do.  And we do this many times every second since every muscle activation is preceded by a low-level, subconscious prediction of its effects.

Three major kinds of reasoning:  deductive, inductive  and abductive.

Abductive reasoning: (Guessing possible causes for an event;  B happened; A might have caused it) QV.

Monica:  key point:  AIs that are generally intelligent will be  ABDUCERS...  predicters... making guesses.

Don't base your images on Spock or Data or the Hal-9000

An AI should be able to browse the web and UNDERSTAND what it reads.

Monica's preferred definition of AI:

"The ability in a computer to easily do most everyday mental tasks that are easy for humans".

(recall Rod Brooks list:

1)  object recognition/ classification of a 2 year old

2)  speech understanding of a 4 year old

3)  motor capabilities  (tying laces) of a 6 year old

4)  social understanding (theory of mind) of an 8 year old.

(That's the real AI program).

Stanley the Stanford DARPA Grand Challenge car can drive on its own but that understanding

is not suited for enjoying the movie  Herbie:  The Love Bug.

Brains are forced to use abduction simply because it's the only thing that works at all in our complex mundane reality.

Monica does not think that the SCARY POSSIBILITY of  Super AI is anything to be worried about...

can't happen (in her opinion) cuz they will always be fallible.  We will always be able to unplug it using just

a community human experts + 10 last generation AIs.  (I'm not sure).

My view:  humans will become incr. marginalized and controlled as  computers comprise a larger portion of the

Problem Solved: Unfriendly AI Monica Humanity HPlus

Attachments