Zeroth Law of Thinking Machines

A thinking machine or a collection of thinking machines may not knowingly harm humanity or by inaction allow humanity to come to harm.

When we speak of 'humanity' in this context we can mean either the human population as a whole or the well being of the species, Homo Sapiens, as distinct from other species.

The above statement is a grand generalisation and fundamental to the other laws. It could not come into operation unless or until thinking machines provided a major part of the administration of the World. Definition is needed. Such definition would have to be thought through before and while thinking machines are being created or while they evolve.

Thinking machines are unlikely to be all of a kind let alone all equally intelligent. Many lesser thinking machines might never be allowed to operate at the level of this law, but be required to refer to higher authority - a superior thinking machine and/or human intervention.

Science fiction has no shortage of powerful and commanding machines. Some thinking machines might calculate that they could reduce the harm groups of people are doing to each other. To what extent could or should they be allowed to take over the running of the world to do so? That is unlikely to have a simple answer valid for all time.

A high level of command might have other types of impact - for example, a thinking-machine driven bioengineering project to improve the durability of human teeth. The species would be improved, suffering alleviated, but the essence of Homo sapiens would remain unaltered. Many people might approve, but there are sure to be hold outs.