Tim Urban, author of Wait but why, has written a great blog post about Artificial intelligence (AI) and its future impact on humantiy. He is writing a fantastic blog about big life questions, and answers posed by a young, contemporary adult.
What Tim says is that AI will most probably develop much faster than we anticipated and that it will profoundly change the world. He narrows down the development of AI in three phases:
– Artificial Narrow Intelligence (ANI)
– Artificial General Intelligence (AGI)
– Artificial Superintelligence (ASI)
First phase (ANI) is teaching computers to do menial and specialized tasks, as playing chess and web browsing, better than humans. The second phase (AGI) sees the development of AI to the level of human brain, and the third phase (ASI) is going even beyond that. As the development from ANI to AGI will still take some time to develop (but it still be very quick – 5-10 years), he warns that (and I totally agree with him) the jump from AGI to ASI will happen very quickly once we reach AGI. The problem is that humans can’t perceive exponential growth very well. This is why, for example we can’t perceive that 7% interest rate (which seems a small number) means we will double our assets (or debt) in 10 years time.
What I find fascinating is the question of interaction between the physical and virtual world. They are still quite separated, with only a few “bridges” between them. The most common are still the keyboard (real -> virtual) and the computer screen (virtual -> real). There is now the emerging trend of the “internet of things”, which may give the AI enough input points from the real world and “hands” to manipulate the virtual one. But maybe still the ASI computers will just stop caring about the real world…
Just a quick edit – in relation, but not totally related