In his book Superintelligence, Nick Bostrom begins with The Unfinished Fable of the Sparrows.
Once upon a time a group of sparrows decided that it would be nice if they had a big, strong owl to help them do sparrow things. When one sceptical sparrow voiced his concerns over how they, lowly sparrows, could possibly control an owl, another sparrow replied:
"Taming an owl sounds like an exceeding difficult thing to do. It will be difficult enough to find an owl egg. So let us start there. After we have succeeded in raising an owl, then we can think about taking on this other challenge."
And with that, most of the sparrows went off in search of an owl. The remaining sparrows soon realised that learning how to tame an owl wasn't easy, in no small part because they had no owls to practice on. But they pressed on as best they could because, at any moment, "the flock might return with an owl egg before a solution to the control problem had been found."
Humans are the sparrows, in Bostrom's metaphor, and ASI is the owl. As it was for the sparrows, the "control problem" is especially challenging because we may only get one chance at solving it. This is because once an unfriendly superintelligence emerges, it will presumably stymie any attempt we make at stopping it or changing its preferences (e.g. the Terminator series' Skynet). (That's why our own Codebots technology can only create and edit eternal code, and not its own, internal code.)
But should we be worried about AI taking over the world?
Well, it depends. Bots may be beating us at board games and TV game shows, but software bot businesses still have a long way to go in developing AGI, let alone ASI.
Even emerging technologies like Codebots require human pilots.
Your brain is an immense network of about 100 billion neurons connected by around 100 trillion synapses. You have hundreds of trillions of neural pathways in your head. No computer in the world today comes close to matching your brain's horsepower.
The Fujitsu-built K, one of the world's fastest supercomputers, is capable of 8 quadrillion calculations per second, and it took 40 minutes to simulate a second of neural activity
Skynet isn't even remotely possible with today's technology. So no, even the most leading edge bots are not going to take over the world, but they may take your job.