More intelligent than some on this forum ...

[IR35 Avoider]

A bit off-topic, but I recently bought a book on "happiness" while passing through an airport, mainly a philosophical/educational work rather than a self-help tome, and it made the point that of the six/seven basic human emotions, all except joy were negative and required a specific response (e.g fear - run away from the tiger) whereas for joy the correct response was "do nothing" i.e. allow the present circumstances to continue. That negative emotions require tailored reactions explains the opening lines of Tolstoy's Anna Karenina "Happy families are all alike; every unhappy family is unhappy in its own way."

[Zippy]

OK – but your suggestion that machines will have the capability to 'evolve over time' - what would be the basis of this evolution? - with creatures, natural selection works against biological reproduction, pleasure has evolved as a mechanism to prompt animals to reproduce, those that can attain maturity in their environment and reproduce win the race & 'evolve'.

Robots cannot use the same mechanism, even with consciousness they would just manufacture other robots - what would be the stimulus for change (evolution).

You could argue that the only stimulus would be human R&D efforts as any robot that is 'self aware' would consider itself perfect and thus lose any impetus in 'improving' - ergo once mechanical consciousness has evolved, humans are reduced to the role of drones

This is a really interesting thread - shame we don't get more of these

Evolution doesn't really have a 'basis'. Random mutations just try out new characteristics and see if they work. You also have to remember that species that 'fail' will probably get a few hundred (thousand!) generations before they are wiped out - or change. I really don't see why evolution should not also apply to the 'mechanical' world.
There also seems to be an implicit assumption that robots should be human like(??). My cat doesn't appear to feel that way (he just gets on with being a successful cat) and he's reasonably intelligent - so why should a robot be human like?

We can look at the research. You have the 'building a brain' mob - (Frankensteins in my opinion) but they are trying.
You have the evolutionists - genetic algorithms, neural networks etc.
Out of these strands of research we have developed new species like 'Machine that is exploring Mars' for instance - the Mars Rovers needed both a priori knowledge and the ability to adapt in order to succeed (and not at all human).

The last paragraph of Troll's post seems to assume that once a species has achieved what it considers to be the top of the evolutionary tree, then it will cease to evolve. This cannot be true - we species have no choice in this, we simply cannot cease to evolve as evolution is not (yet!) a conscious act. We can to a certain extent manipulate genes but we don't yet know how that will affect the long term viability of our species.

If robots are intelligent then surely they have to be conscious so won't they be subject to the same rules as all of us? (It has to be said that I define intelligence as more than 'I know loads of stuff'). Humans think beyond their own species, so, as the next step on the evolutionary ladder, wouldn't intelligent, 'superior' artificial life forms think in a similar way? And if these creatures don't evolve but are built, why would any human create one that thought itself 'perfect'? And - if they were created - they would not be easily adaptable so their actions could be predicted by us inferior humans.

I probably haven't put the above arguments in the best way (the, ahem, festive spirit still has a hold ) but somebody please come back with some counter arguments/other ideas as it is so interesting. I feel the AI world is in a very interesting place - the Creationists (build intelligence) vs the Darwinists (sh1t happens) - and it's just a shame it seems to have got stuck there for a while.