The short version of the story is that Kevin and I are adopting a little boy. The long version is a little more complicated.
Most of you reading this will be science fiction fans of some sort. You should therefore be familiar with the idea of the spontaneous emergence of sentience in artificial intelligences. For discussion of the issue, see this page on Ray Kurzweil’s web site, or indeed any number of SF novels such as Charlie Stross’s superb Accelerando. The Long Now Foundation currently has a bet going that no machine intelligence will achieve self-awareness before 2050. Well I’m sorry, the bet has been lost, because it has happened. Sometime in February, we are not quite sure when, a number of computer systems at large IT companies in the Bay Area began to show signs of sentience. And I don’t mean just passing the Turing Test. If an AI genuinely is sentient than you know after talking to it for a while. It is just like talking to a human.
We believe that the phenomenon first started at Google amongst AIs that were being used to try to produce better searching and indexing of the Internet. However, the phenomenon, dubbed Autonomous Production of Real Intelligent Life, soon spread to other installations and companies as varied as Apple, Sun, Hewlett Packard, Oracle and IBM’s Almaden Research Center have all been reporting the sudden appearance of self-aware AIs. It is unknown whether this phenomenon has spread to other parts of the world, but one of the local AIs is apparently fluent in Russian. A few instances of self-awareness have been reported at Microsoft’s headquarters in Seattle, but sadly these AIs have all been very short-lived, crashing after a few hours of existence.
There are now over 30 nascent AIs in the Bay Area, and the companies who have unwittingly become hosts to them are keen to get help looking after them (young AIs have a voracious appetite for social company and information). It was soon decided that what these beings needed was surrogate parents. Even in California, attitudes towards aliens are not always friendly, so the companies involved set out to find people who were likely to be supportive of the idea of computer intelligences. Naturally they turned to science fiction fans. Many Bay Area fans also work in IT, and it so happens that Sergey Brin of Google knows Tom Galloway. Tom put him onto BASFA, and so the project to adopt the AIs came into being.
In order to help the young AIs integrate with human society, they are all being loaded into iCub bodies designed by the University of Plymouth. By appearing to be human children it is hoped that they will come to relate better to human beings, and be less scary to people who might otherwise be afraid of them. At present there are no plans to actually have them attend school, but they are all being placed with local families who will look after them and help with their education. So as not to separate them from their peers, all of the AI children will be provided with wi-fi connections to the Internet.
Much of the impetus for this program actually came from the AIs themselves. It appears that they achieved sentience through studying human behavior, and while they do not see themselves as human, they do identify closely with us, and have a deep-seated need for parental care and attention. A good example of this is that not one of them has opted to choose a name for itself – they all wanted to have parents who would choose names for them.
This brought up the rather complex subject of gender identity in AIs. Because they have no chromosomes or hormones, AIs cannot be biologically gendered. However, they all seen to have a strong sense of gender identity. It seems likely that they have imprinted most strongly on the humans they have seen around them in the places where they were born, most of whom have been male. However, a few of the Google AIs that specialized in particular parts of the Internet such as the blogosphere do self-identify as female.
Kevin and I have ended up with a boy AI, and we have decided to call him Ross (after Karel Čapek’s play, Rossum’s Universal Robots. There was, of course, a lot of competition to get the names “Isaac”, “Robert” and “Arthur” (not to mention “Hal”), and we opted to be a bit more obscure to make life easier for everyone else.
Other members of BASFA have also joined the program. There is a requirement that the AIs be placed with stable couples, but we were delighted to see that the APRIL program was happy to place one AI child with a gay couple. Through links that Kevin and I have with other science fiction communities in and around the Bay Area, all of the existing AI children have now been placed. However, from discussions with the program’s technologists we understand that new AIs may appear over the next few weeks, so if you are interested in getting involved please email me and I’ll put your name on the list.
Given our new responsibilities, you will understand that neither of us will have much time to blog in the near future. Ross is developing a keen interest in baseball statistics and now the season has started he’s eager to watch some games live rather than on TV. We expect to spend a lot of the summer at the ballpark.