Hacker News new | ask | show | jobs
by yaacov 3653 days ago
I can't tell what this is supposed to mean. Taxing employers for using robots as if they were human employees might be a good idea. But giving robots 'rights and obligations' as if they were human is totally insane.
3 comments

They just want taxes and the ability to control robots "replacing" workers.
Why tax companies for becoming more productive though. Especially considering that international competitors which are not taxed would take their market share. People often don't realize that the company won't necessarily make more money once it starts using robots - in competetive markets, profits approach zero.

The real problem with automation is that it might (or not) increase the inequality in the value of human labour. The solution is income redistribution.

I can see a case for obligations much more than I can see one for rights.
Well, someday we might inevitably need to discuss robot rights, but that'd be somewhere past the point where we can declare a robot a sentient being. ;)
Why is being 'sentient' (a vague term) a requirement for something to have rights?
I'm certain in this context "rights" refers to the rights regularly afforded to human beings (as in the european proposal), not to the rights we might assign to non-sentient animals.

In any case, it seems to me that we don't understand consciousness and life enough to truly define any requirement for rights any better than the meaning of 'sentience'.

Until it is sentient, it's just a hammer with pre-programmed response. After that, it can define itself.
So, sentient means 'having the ability to learn new things' ? Kinda like humans, who are born with certain instincts, but are then able to learn new things or 're-program' themselves.

Well, there is a filed called 'machine learning', which helps design robots who will be able to learn. If they are able to learn and exhibit intelligent behavior, then we have to treat them as something more than rocks or hammers.

That is what the document is referring to - not pre-programmed robots, but AIs (robots who are able to learn, perceive, change behavior etc.).

> it can define itself.

I think you might have missed the full implication of that part.

Certainly, a sentient being can learn new things, but that's a matter of rule-base and memory. No actual intelligence required, IMO. Sentience, to me, goes toward a concept of self. Of being aware that one exists and is responsible for one's own actions.

Personally, I fine with having really smart-but-not-sentient hammers. I'd rather not have to lard up machinery with some touchy-feely "it's alive" sentiment. We won't have to have lots of hand-waving over whether or not that hammer is a slave or not.

If we can have an army of dumb robots performing the work that slaves do now more efficiently and cheaply, then we can easily spot the slave owners and remove them.