Should we be worried about 'killer' AI robots?

"However, speaking to BBC Three, Professor Peter Asaro, vice chair of the International Committee for Robot Arms Control, shared his concerns about battlefield AI and wonders if it could ever be used against civilians.

"The issue is that computers aren’t capable of determining when it’s legally or morally acceptable to take human life, and can’t be held morally or legally responsible for doing so. 

“We’re primarily focused on banning these weapons for military use, but they could proliferate into other areas of civilian life. A particular concern is if for example if they are adopted by police forces for crowd control, or by private security forces for securing buildings.”"