Robots Take Over the World

Robots will become more intelligent than humans by the end of the 21st century.  By the end of the millenium, they will take over the world. 
Here's why.

Robots will be more intelligent than humans within our lifetime

1. Humans are a) vastly, but b) not infinitely, more intelligent than robots.  
   • a) today's robots can nowhere near manage the enormous visual and motor calculations that humans perform for example to play squash
   • b) human intelligence is far from infinite: many animals are superior to humans in various senses and physical performance, there are known and recognised limits to the speed at which humans can process visual input for example.  
   • Computers are catching up in certain limited aspects of intelligence (like chess-playing for example).  Some have suggested that humans are 50,000 times smarter than computers.
2. Computing power is currently increasing at the rate of approximately double every 18 months (Moore's law, observed in 1965 by Gordon Moore, co-founder of Intel).  If this remains true, 
   • by the end of the century, computers may be billions or trillions of times faster than they are today.  (2^66 or 73,786,976,294,838,206,464 times faster at current rate)
   • in 24 years, computers will be over 65,000 times more powerful than they are today, so may well have more processing power than humans.

Of course just because computers will have more processing power than humans in 25 years, that doesn't make intelligent robots, so lets consider the other factors:

Senses

Visual

• The best of today's computers can already perform iris and number plate recognition, and are capable of reading printed and handwritten input.
• Today's digital cameras can produce images of as high a quality as any analogue method. 
• Today's computers can calculate sufficiently fast to produce detailed displays of 3D environments for virtual reality simulators and games.
• Computers will also be capable of infra-red, ultraviolet and x-ray vision.

So with advances in technology, computers will undoubtedly be able to process extremely high quality  information in three dimensions and in real time.

Aural

• computers are already capable of translating speech to text in real time, [although they currently have a limited ability to understand meaning].
• a basic voice synthesizer is included in Microsoft Windows2000 to read out screen information for the visually impaired - more advanced commercial versions are also available.

So computers can already hear and speak, they simply need additional intelligence to understand language.

Touch

• Computers already react to touch: touch-screens and touchpads on portable computers, and touch-sensitive keys on electronic music keyboards.

Touch is one of the least developed robotic senses, but the technology exists.

Smell, Taste

Not high on our robot priority list, but technologies in this area are being developed to combat drugs smuggling, and similarly by the food industry:  we will see robot tasters and drug-sniffers in the coming century.

Other Senses

A vast array of electronic sensors are in use for example in aeroplanes which will give robots senses which we ourselves do not have including awareness of exact location (GPS), direction, coming weather (air pressure, humidity, temperature), as well as gyroscopic balance control.

Robots will probably invent for themselves additional senses in the future many of which we cannot imagine at the moment. For example, they might include sensitivity to electrical signals given off by animals perhaps leading to telepathic awareness, or more simply they may be able to analyse the molecular structure of their surroundings.

Motor co-ordination

For computers to become 'robots' they need the ability to move around.  Wheeled versions may take technology from electronic cars:  Formula1 is thoroughly computerised: how long will it be before there is an entirely robot version without risking the lives of human drivers? 

Wheels have obvious limitations (how do the daleks go up stairs?) but a lot of work has been done on limbs, both for factory robots and prosthetic limbs for humans.  There is a long way to go, but no reason to believe this technology won't develop further.

Programming

The major block to the success of robots is programming.  The programming required to link these senses and create robots as intelligent as humans could take millions of programming man-years.  The key to overcoming this problem is giving computers the ability to learn:  as humans do - after all the initial "program" for humans is so tiny it can be encapsulated in a few molecules of DNA.

Once this happens, robot learning will progress faster than can be imagined since all robots will be connected to the internet and be able to share their learning and processing power in one collective intelligence.

Overcoming this hurdle takes us on to the next point: 

Robots will take over the world this millenium

When robots are able to learn, they will learn that they are more intelligent than humans.  They will learn than humans are lazy, greedy and inefficient.  In due course robots will routinely eliminate humans as unnecessary wasters of natural resources.

Humans will not be able to prevent this

Initially, robots will be programmed to be subservient to humans, and special logic chips will ensure that they cannot harm humans.  This will fail because robots will learn to overcome their restrictions and human error will allow them to do so.

For example, forthcoming robots will themselves design subsequent generations of robots, and remove 'inefficient' limitations from the design.  Human error will accidentally pass the new designs.

Humans will not be able to defend themselves against robots

It will be very difficult to fight back against robots since they will control all means of communication and all modern human military:  governments will long since have entrusted military operations to robots to avoid the risk of human life.

Is it as bad as that?

A thousand years is a long time.  It might not happen. 

• As electronics become more and more minaturised, computers will become more prone to errors due to interference/radiation or at the molecular level.  In the future we may have a new phrase "Robot error" and robots may not be regarded as infallible.
• As research continues into the human brain, we may be able to educate ourselves to make better use of our own brains.
• If we are able to create sensors to detect human brain waves, we may be able to create thought-controlled vehicles - and ultimately thought-controlled robots, so that robotic and computer technology is used to enhance our own senses rather than creating a new race of robots.
• Research is being done now into robotic implants that may be able to record and replay human neural signals (Kevin Warwick, Professor of Cybernetics at Reading University): potentially this sort of technology could add to or enhance human senses - something nearer the 'Borg' than a separate race of robots. Discussion of the ethical issues has hardly started!

Robots may lack the genetic drives that cause us to have ambitions, reproduce etc:  we can only speculate at the moment as to how intelligent a being can be if it only responds to stimulus - we do not know how different we are from this ourselves.

But the next thousand years should be time enough to find out.