Bletchley ParkThe top-secret base of English codebreakers during World War Two. , where the first proper computer was unveiled in 1944, will be the meeting place for the delegates at the first Global Summit on AI Safety.

Among them will be two of the three men nicknamed "the godfathers of AI": Professors Yoshua Bengio and Yann LeCun. Bengio believes that AI is as big a threat to humanity as pandemics and nuclear war.¹

But LeCun calls these fears preposterousRidiculous. . "Intelligence has nothing to do with a desire to dominate," he argues.²

These are the top 10 risks that AI poses: 

Tech takeover It could become smarter than humans and start running the world without us. If it saw humans as obstacles, it could set about destroying us. 

Threatening weapons AI could be used in an arms race to design new weapons, such as drones that decide for themselves whom to kill.

Terrorist risks Illegal organisations, or even mad individuals, could use AI to attack people or critical infrastructureThe basic systems and services that a country needs to keep it going..

Election misdirection AI could stir up unrest and undermine democracy by spreading false news or creating deep fakes.

Automation trepidation AI could be used instead of people to carry out many jobs, creating mass unemployment.  

Overdependence tendency We could become dependent on AI systems which are too complicated for humans to understand. Then, if things went badly wrong, we would be unable to put them right.

AI spies Smart devices could relay private information about their users to the manufacturers. There are already worries that smart speakers may be eavesdropping in homes that have them. 

Deep-dyed bias AI could be trained to favour people from some groups over others.  

Censorship grip AI could wipe the internet of information on certain subjects and block whole sites.

Crime time Criminals could use AI to mount ransomware attacks, hack into bank accounts and organise scams.

Yes: AI is developing at extraordinary speed. The many different ways it is designed mean that we cannot predict what it will do. One expert points out that it is less regulated than a sandwich shop.

No: According to Yann LeCun, AI is nowhere near posing a threat: "Until we have a basic design for even dog-level AI (let alone human level), discussing how to make it safe is premature."

Or... It is no bad thing if they are. Once humans realise they are in danger of being wiped out, they will have no choice but to behave as well as possible in order to convince the robots of their worth.