Rise of the Machine Rights
I’m in a book.
The course I took last year finally materialized into physical from couple of months ago. I’ve no idea if this book is actually available anywhere, even in a digital form. Well, at least I got mine.
In the book a group of doctorate students from three universities in Helsinki wrote cross-disciplinary visions of what certain IT innovations will break through by 2030.
The future is notoriously difficult to predict and the future of technology even more so. So, I’m pretty sure that any predictions we have made for 2030 are going to be wrong. There are some things we can be pretty sure about and try build on them, though. For example, technology will get better. The western population will grow older. Fusion energy will always be here in 30 years.
My second group wrote about intelligent or smart machines on how we see that there are some non-technical barriers that have to be broken before we can see robots and machines everywhere. Some of our ideas are also presented at the 26C3 conference under the title “Here be Electric Dragons: Preparing for the Emancipation of Machines“. So, if you’re in Berlin around Christmas, go and listen to our fantastic ideas. Unfortunately I can’t make it there, but my co-author Lorenz Lechner will be there to entertain the audience.
One of our core ideas is that for autonomous (or, as we put it, ultimate) machines need rights. One reason for this is that normal product liability is not enough if these machines have AI dictating their decisions. If giving rights to machines sounds strange, it shouldn’t. In a sense this is comparable to human rights and the idea of corporation as a legal entity, where the corporation and not the shareholders are legally liable for its actions.
The follow-up question is of course how to manage the risks that autonomous robots pose? We are pretty good at managing all kinds of risks. One approach is to design fail-safe systems. The other is using insurance.
One of the challenges is that the machines of tomorrow and even today are more and more dependable on the software. We can’t end up in a similar situation with robots as we have done with commercial software – no guarantees whatsoever (see for example the capitalized(!) section 17 of a Microsoft EULA or similar section 16 of GPL). We believe that through an insurance of sorts, these sections could be shortened two capitalized words, DON’T PANIC, instead. Preferably in large, friendly letters.
Also, we believe these issues are urgent. The actuality of the technological development was really nicely illustrated by a recent xkcd comic (note the mouseover text). We have also discussed about this subject previously on this blog, Vincent already wrote about the relationship between man and machine early this year.
If you can’t make it to 26C3, here’s a copy of our paper “Augmenting Man”. We are currently in process of refining it and trying to pimp it for other publications.
LikeRelated posts:
