I start with mindfiles and mindware. Mindfile is a dump of the contents of a mind, perhaps mine, and mindware is the software which uses the mindfile to pretend to be the originator of the mindfile. The author, described as a lawyer, technologist and medical ethicist, has paid for the construction of BINA48, a stab at something of this sort, using a mindfile built for her partner, Bina. She says that one of Bina's early comments about BINA48 was to complain about the latter's choice of blouse. Lots of other people are beavering away at other things of the same sort, not least the people who make elaborate games to be played in elaborate worlds simulated on computers.
One place to start a mindfile would be someone's email, facebook and twitter accounts, plus the logs of their activity on their computers and telephones. Someone active in all of these domains is, in the ordinary course of events, building up a lot of stuff about themselves, just waiting to be harvested. A more deliberate extension would be to wear something which captured everything one saw, heard or said, this last being available for just a few pounds, now.
The author believes that lots of robots built with such mindfiles and mindware will built in the years to come, that some of them are going to be, in many ways, indistinguishable from humans and will, in particular, be conscious in more or less the same way that we are. All this in the next twenty years or so. I believe that this is likely enough that we should not leave the thinking about how we might respond to the writers of science fiction, writers who seem to have been in the lead so far. Certainly there are plenty of quotes from such in the book.
One of the drivers is the quest for immortality. Some people rather like the idea that when their flesh wears out, they have some chips ready to take over. I wonder whether a computer replica of me would suffer from comparable problems of aging, that it would turn out that something clever enough to be like me would turn out to be so like me that it suffered from some of the same problems. That the part of aging that is to do with the decay of the chips can be dealt with, one just slots a new lot in, but another part is to do with the accumulating muddle and dysfunction in the huge amount of data running around on those chips, muddle and dysfunction which might prove much harder to deal with, while preserving any of the robot's sense of continuous self. Perhaps we would have to ask the robot whether it wanted to take a chance on the latest upgrade, or the latest spring cleaning, rather as a doctor would ask a real person whether he wanted to take a chance on having his heart patched up.
I associate to the way that software evolves over time, in some ways not unlike the much slower evolution of, say, mammals. Because, like mammals, software is built largely incrementally with all kinds of new stuff layered on to the old stuff. Things get left behind, like spleens. Other things get re-purposed. Some things acquire extra functions, not needed or not thought of at the time of their first appearance. One can get into the position where a chunk of software has become so messy that it has become very difficult to change or even to mend it. But most of the time it continues to do a good job and it would cost a fortune, possibly a large one rather than a small one, to start over. Just think of what would be involved in building a replica of MS Word, from scratch, without access to any materials from the existing Word, other than those which can be derived from using it.
But I digress. The book contains lots of stuff about how we might respond to these robots, in particular what human rights they might acquire. The author points to analogies in the way in which various groups, for example women, have slowly acquired full rights. She points out there is another precedent in that non-humans such as companies and charities already have legal existence. Could I endow my robotic replica with my property, with full powers to dispose of it, or not, as it chose? That the robot could be regarded as a sort of animated will? Or function as a sort of supervisor on the sideboard of his or her heirs?
She makes some important points with which I agree. For example, she believes that free-will might be a bit of a con, a deception, but one that it is, nevertheless, a convenient and comfortable fiction. She believes that top-down rules of behaviour are not going to be enough, are not going to work well enough. Such rules needed to be complemented, if not replaced, by statistical analysis of what has happened, what has been said in the past, analysis which will come up with probabilities rather than certainties, but which will still do a better job than the rules. See the post at reference 2 about my road to Damascus moment last year.
Another part of her response to the rules problem is the idea that there will be panels of experts empanelled & empowered to rule on all kinds of matters robot. A panel to review, for example, whether this or that brand of mindware is good enough to be allowed out in robots on the streets. A panel to review whether this or that brand of mindware qualifies its wearers as citizens. Lots of panels, a large scale bureaucratic version of the famous Turing test of computer intelligence (see google in general or reference 3 in particular).
The book also contains lots of what I might call California speak, despite the fact that the author comes from Vermont. All kinds of bubbly stuff about life and the universe which used to froth up from minds sauced up with recreational drugs when I was little. Some of this was tedious and I admit to some skimming. Some of it reflected the dangers and difficulties confronted by an able person tramping around fields other than her own; but then again, someone has to do it, to get the debate started, and she may well have succeeded at that.
I was left rather concerned. That these robots were going to come and that we would not respond to the opportunities and challenges they were going to present very well. That, certainly in the early days, there was going to be a lot of mess, muddle and worse. Perhaps they really would take over the world, in the way suggested in many science fiction films, films which usually pass us by as they usually seem to involve large metallic monsters noisily charging around and smashing up cities - until, that is, the current version of Arnold Schwarzenegger can get to and zap them. Or perhaps the man himself, taking a break from his retirement from political activities.
Reference 1: http://www.psmv2.blogspot.co.uk/2015/01/virtually-human-1.html.
Reference 2: http://psmv2.blogspot.co.uk/2014/09/och.html.
Reference 3: http://www.turing.org.uk/scrapbook/test.html.
No comments:
Post a Comment