Some of the Google searches I have done recently turn up mostly automatically written articles that plug in numbers and facts but make no sense at all, because there is no real understanding. (Yes, this is probably not "real" AI, but it's the sort of thing people are and will be trying to apply AI to)
It seems to me that there is no way to get to a serene, well-functioning society that utilizes intelligent machines, because in the process of developing software that can digest and write intelligently about topics humans are interested in, we will inevitably come up with less intelligent software first, that produces plausible misinformation far cheaper than humans. And economics will drive out, is already driving out, real research and journalism. Human society would be drowned by automatically generated misinformation by the time a machine is truly intelligent, whether that is in a couple years or a decade or longer.
I'm increasingly thinking that the people who are worried about artificial intelligence ending the world as we know it have slightly missed the mark - superhuman intelligence is a singularity that we are approaching, but we are going to be ripped apart by the tides beforehand and maybe that will prevent the endpoint from ever happening.
No comments yet.
