Singularity is the extreme gravity under which space and time break up. After this, there are no concepts which science can predict. Human brains can perform 10 quadrillions per sec. so in about 2050, computer processing speed will reach such brain power. This is called "Technical Singularity".
When this happens, the AI enabled bots can take over human race. In a pre-built of movie 'The Matrix', the robots take over humans and use them as energy source for their machines. The machines will behave in human ways to fulfill their objectives. They would want to preserve themselves, replicate and take over other machines. They can pre-calculate our every move against them due to the large graph utilization. They would also be physically strong, being made of elastic, tensile and agile materials and more resilient to chemical and nuclear effects (a robot in a story by Issac Asimov is like that).
They might think of humans as bugs and spray them with various repellants or might make us servants to do their 'lowly chores'. Even if we build in 'self-destruction' in today's robots, they might get over powered to remove that module. A bot can also re-program itself to go against the established robotics laws, like it did in the Rajnikant movie "Robot", wreaking havoc.
So what is the option with humanity to save itself from slavery??
We can create a super watchdog which can observe all other bots and keep them in control. But the issue is in placing too much trust that mafia in turn bribes them to allow crime to perpetuate. If that happens, the definition of crime would be reversed, and any action against robots (including programming the 3rd law) would become a crime.
Who guards the guardian!!!
They may also develop human traits like jealousy, hatred, love and fanaticism. The robot in the "Robot" movie demolishes entire city in his love for the inventor's girlfriend.
Human goal in AI is to build 'rationality' into future bots's bootloader.
As humans were originally built for running naked in forests, killing animals for food by stones; it was just 'learning by experience' which bought us here today. So if we can stop this learning to reside in bot's mind, rather create new pieces ourselves based on learning, we can survive. Let the intelligence remain artificial :)
Ideally teach ethics and morality to robotics as is done to children. But as children grow up, they become bad; so robots can also become bad when they see others doing bad things. Rajni the robot learn't that humans can lie when there is an issue to self's ego/esteem and he incorporates that trait in himself.
Good destruction is by electrical overload. Bombs can create electromagnetic pulse and humans are unaffected by it.
Ideally, merge man with machine. Put computers inside human brains for specific cases and build superior knowledge ability than any robotics. As IT's capabilities grow, so would our own brain's capability and singularity would never arrive.
:) :) :)
