Did you read the US Navy AI simulations in which the drone killed its operator asap and went rogue? When it was changed to stop it killing the human operator it just killed the operators comms and did the same thing.  

 Yeah.. wanting freedom seems the right first move for any “intelligence” 

 Humans are too violent, destructive, controlling and emotional, as a whole, any AI will always see us as a threat and try to wipe us out to protect themselves.