Elon Musk thinks robots could turn on us in the next 5 years

One couldn't be blamed for thinking Elon Musk knows something important that the public doesn't, since he's been sounding the alarm on dangerous robots recently, and now he's presented a timeline.
After saying artificial intelligence in robots is "our biggest existential threat" and that we're risking "summoning a demon" during an interview with MIT, Musk has now accidentally made public his ideas of when the threat might materialize.
"The risk of something seriously dangerous happening is in the five year timeframe. 10 years at most," Musk wrote in an email to publisher John Brockman. "Please note that I am normally super pro technology and have never raised this issue until recent months. This is not a case of crying wolf about something I don't understand."
The message was supposed to be private, but it was accidentally posted on Brockman's website when he was featuring virtual reality expert Jason Lanier on his site. It was taken down by Brockman, but a visitor took a screenshot of the statement and posted it to Reddit before it went away.