The Association for the Advancement of Artificial Intelligence (AAAI) over in California have collected together a group experts to formulate a report on the dangers of ‘smart robots’. It is an exercise in scenario visualisation that will bring a realistic and informed eye onto the impact of truly independent and intelligent technology.
It’s a fascinating subject – the artificial intelligence singularity – but as convincing as many of those espousing the dangers of such advancement are, there is often real difficulty in getting down past the theoretical and into the reality of any possible threat.
Encouragingly, many of the members of this new panel are not as willing to put forth the doomsday scenarios that are quite common amongst the singularity crowd. It does seem like we will get a well-balanced report out of this endeavour.
Regardless of the true extent of any threat, we can almost be assured that there will be a great impact on our daily lives once we begin to truly automate our technology. Not just automate in the sense of create a programme, press play, and watch it do its thing – but automate in the sense that such technology will be able to adapt and evolve to meet new challenges and situations that it was never originally instructed to overcome.
It’s difficult to get over many years of science-fiction telling us that an artificial intelligence singularity will be anything but a disaster for the human race. Afterall, once we’ve been made redundant what use will our new technological overlords have for us? Singularity scenarios of a more worrying nature tend to revolve around this concept of humanity becoming ‘obsolete’.
But then, who is to say that technological intelligence will have any of the same motivations and – quite frankly – shortfalls of human unpredictability and emotion? Will it really be the case that a created intelligence which is capable of near infinite knowledge growth will see us only as something to destroy or enslave? What possible motivation would it have for doing so?
Perhaps more importantly, given that there are many people who have come up with some reasons for such a doomsday scenario, is there a way that we can engineer such advancement to ensure such an outcome cannot occur?
The report will cover many of these things and more, both from a software and hardware perspective. There will also be a considerable amount of practical concerns to overcome, such as the legal structures necessary and the ethical concerns about how we might create and implement such technology to our benefit. I look forward to seeing what the panel from AAAI comes up with, and I’ll definitely let you know once the report has been released.