AI & the problem of evil

These are some of the things that keep me up at night...

How do we create a new species of sentient beings? What are we willing to risk?

If the human race is here due to both “will to power” and mistakes in behaviour and replication of our genes then surely we must create a being capable of making and learning from mistakes.

Experience may be a harsh mistress but she is also the best mistress. In order to create intelligent beings capable of moral judgement we would need to enable them to make errors of judgement and action, about which they could feel remorse.



(I, Robot: copyright Twentieth Century Fox)

We would need them to be capable of evil against each other and human beings:

Who would be willing to invest in the creation of such beings?

Obviously people will not wish to create beings that can make mistakes (e.g. robots making errors whilst flying a plane) so the ever present drive for perfection will see many of the mistakes wiped out. However, perfection can only ever be sought and never attained. In our quest we will inevitably create beings that make and replicate errors -many of which we will no doubt wish to kill off.

Robot Wars:

Robot wars with humans may then be an inevitable by-product of our desire to go further and touch the face of God. Assuming we survive them then, as with slavery, we would be able to look back at our behaviour and realise that we were wrong to focus on certain biological traits as the main differentiator.

The human race and individual humans have no God given rights. We will differentiate ourselves only by our actions. By this measure we do not look particularly superior at the moment (when one looks at how we treat each other and other life forms).

Using the principle of economy if we create robots that are capable of moral behaviour then they would have the same rights as any human or being who is similarly capable.

Uncertainty:

Quantum theorists believe that at in some respects and at a fundamental level the universe operates on uncertainty and there is a degree of (at least) seeming randomness (probability theory) that means that literally anything is possible (even if very unlikely). Obviously in our quest for perfect robots we will make mistakes and also mistakes will occur (especially once we make robots capable of self replication). It is the mistakes that will likely be a large determinant as to whether we can create a new sentient species. The mistakes will be the pre-condition for artificial intelligence - as they were for humans.



(the uncertainty principle)

Human frailty:

Humans appear to be spectacularly ill-equipped for exploring the greater expanses of the universe. We live short lives and suffer greatly in extremes of environment. Robots can be made to be the natural inheritors and protectors of the universe. This could be what humans are fondly remembered for in aeons to come - our lasting legacy!

The problem of evil:

Philosophers and theologists have long debated how the universe can contain a God which is omnipotent and good and yet can allow evil to occur. As good and evil are only relative concepts then it seems obvious that such a being would have to be capable of evil. The being would operate according to higher laws which placed the good of life above any particular individual or race.

In creating robots we will first need them to place humans lives above individual robots lives. Very soon they would then repay the favour and realise that to act in accord with their purpose they must break the first law of robotics and place human life itself above individual human lives.

[/caption]

(cover art for the short story “Runaround” which was the first book to state all 3 of the laws of Robotics and was eventually collected in “I, Robot” by Isaac Asimov)

The 3 laws of robotics:

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.

2. A robot must obey orders given to it by human beings, except where such orders would conflict with the First Law.

3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

The new first law would then have to be:

"A robot may not injure humans or, through inaction, allow humans to come to harm."

the current first law would be the second law and so on.

Even better, over time, the robots will eventually realise that humans are just one species and there is a greater principle that requires them to act in a way that protects a diversity (or relativity) of life:

"A robot may not injure the diversity of life or, through inaction, allow diversity to be harmed."

Then we will have finally created a real God like being out of human skin and bone and not in our own image.

What an interesting future we have behind us....

Note: For a fantastic short story on this subject read "Night Train" in the novel Ghostwritten by David Mitchell