Clearly artificial intelligence will upset the world through its applications in different areas:

Autonomous Vehicles: Ford recently announced that they will have a 100% autonomous vehicle on the market by 2020.
Robotics: More and more jobs are at risk of being removed and replaced by robots such as front desk and reception jobs. The  Pepper robot that uses IBM Watson technology, which can arguably be  described as used in banking agencies in Japan, and more recently  announced at RATP in France. For now, it's just a stammer.
Natural Language Processing: This will revolutionize our interactions with computer systems. No more need for screen, keyboard and mouse. The  computer of the future will look more and more like Google Home or  Amazon Echo, which I already find very good, even if they are just  beginning.
IoT and Big Data: The Internet of Connected Objects is still in its infancy. More  and more objects will be connected and will be able to relay a mass of  information either to a human or to a control center for decision  support, which itself could be automated. For example, if your car transmits real-time data on the level of vibration on your brakes. There  may be a decision to notify your garage to check your brakes or even to  stock the necessary parts when you come for your next maintenance  (preventive maintenance) or decision making to immobilize the vehicle  safely in real time before disaster. This already exists but will be done on vehicles in mass production. This applies from the car, through the trains to the dishwasher.

All  this will pose interesting ethical problems: When an autonomous vehicle  will have to make the decision to sacrifice its driver against the  pedestrian who crosses in the event of an imminent accident, which  decision is the good one?