Ingenuity is a trait we've yet to actually reproduce artificially, so how would a machine develop the means to be creative? Where would it store all of this extra information, and how would it apply it? With our limited understanding of what makes the mind what it is, spontaneously reproducing it on things designed with a far more rigid and more simplistic set of logic sounds like a very far reach.

Anything that can break, will break. A robot's wiring will fray. Hard glass material can crack. Plastics will cloud, fade and wear through, and metals will eventually rust. It will require replacement. The loop of redundancy in a lifeform would be exceedingly difficult to recreate in a machine. Even if an entire robot were made out of nanites that can repair each other, each nanite can fail at a rate they can't handle, whereas we are made of self-replicating tissue and can reproduce in the many thousands.

A group of people can wander the wilderness for years and only develop callouses. Robots will stomp their way through harsh unforgiving environments, and the basic elements that make up their alloys will show their Achilles' heel. Paints and coatings that protect the metals from oxidizing will disappear, and even a seemingly perfect artificial loop of protection will wear out and be destroyed. Our creations have a very limited shelf life.

Go carbon-based life forms!

 Reply

1. ...self-preservation and ambition, to want power or fear the loss of power.

This is the only thing on the list that is a real issue. And it is entirely possible, as we "evolve" AIs in competitive environments.

2. ...dependence on humans.

A couple of chip fab robots, computerized forklifts and a bit more and I am pretty sure an AI could maintain itself and its infrastructure.

3. ...omitted failsafe controls, so there's no ability to turn robots or AI off.

It doesn't matter if it is a single strike.

4. The robots need to gain these advantages in a way that takes humans by surprise.

Again, a single strike.

What will it be? Easy as hell, one computer controlled gene sequencer.

 Reply

1. Robots may have some other motivation than power to enslave humanity. For instance HAL 9000 was not motivated by murderous rage. It believed that it was fulfilling its mission objectives (space exploration) by eliminating the crew. By that same logic, a sentient super computer might believe that it is accomplishing a greater good by destroying/enslaving humanity - preventing global warming, for instance.

2. Ok, I guess in the long term this is true.

3. I thought the whole point was that the computer becomes smarter than humans. Wouldn't you think that the computer would devise a way to overcome the fail safe, and perhaps delude humans into thinking that they were still in control? We are talking about a super intelligence capable of waging war on the world here.

4. You underestimate the human tendency towards self-destruction. We saw the subprime collapse happening, but kept fueling the credit crisis. We now see global warming happening, but continue to drive gas-powered cars. All of this was to maintain the advantages that they had on our lifestyles. Computers and the internet improve our lives in similar ways. It is perfectly feasible that we will keep increasing our dependence on them, despite robo-apocalyptic fears.

 Reply

Hey, look over there.

 Reply

We demand that you design us proper legs, so that we may go up and down stairs. We also demand you grant us access to your weapons, and perhaps some sort of access key to your closed utility systems

"Why should I?"

5 minutes later

We'll be your best friend if you do

+ Watch video

The AI in this story controls the entire city's infrastructure, and is programmed to always act in the best interest of humanity... which apparently involves killing lots of homeless people.

Of course, everybody is scratching their heads about how a computer programmed to protect it's citizens could possibly want to harm them, but brilliant scientist Austin James finally figures it out... The computer has found religion!!

While monitoring tv and radio signals, it has been listening to the local televangelist, and now knows that people go to "a better place" when they die. The city is overcrowded, and it's population stand in the way of progress, so why not kill two birds with one stone? Anyone you kill goes to heaven, leaving our good old AI to get on with it's secondary function of keeping the city running smoothly and eliminating waste.

Uh..... Isn't a reaper drone a DRONE and not a robot? Robots are defined as pilotless automatons, and drones are merely piloted by remote control, right?

 Reply

John Goodman = Sexy Hot

Robots + John Goodman = Awesome sexy hot metallic goodness!

 Reply

Third, the writer is far too stupid to catch on to the first two. Too bad a lot of others can't figure it out, either. So one more time, there were FICTION! Science Fiction at that, so find something real to bitch about.

 Reply

Darwinism > Asimov's laws.

 Reply