Recent AI (artificial intelligence) safety concerns have been fuelled by stark warnings from AI scientists about the growing risks of the technology going rogue.

Some of the most pessimistic predictions warn that there is a greater than 10% chance that AI could wipe out all humans in the next 10 years.

Many of the apocalyptic messages come from observers of the AI world known as 'AI Doomers' - but the big question many are asking, is how these Doomers think AI could actually bring an end to mankind.

Cyber Correspondent Joe Tidy looks at some of the scenarios being imagined and assesses how likely they are.