What could go wrong with killer robots?

Autonomous killer robots? They're not just for science-fiction films any more.

Powered by automated translation

The concept of "human rights" takes on a whole new dimension next week when the UN Human Right Council takes up the issue of autonomous killer robots.

Don't laugh. It may sound like science fiction, but last week Dmitry Rogozin, Russia's deputy prime minister, revealed that his government is developing "terrorist-killer robots" capable of identifying and attacking targets without human direction.

The idea, Mr Rogozin said, is to be able to respond to terrorists after major attacks, in cities for example, with minimum human casualties. Put that way it sounds more or less reasonable, but once the technology exists … well, it is too easy to imagine "terrorist-killer robots" turning into "terrorist, killer robots" in the wrong hands.

And then there are the malfunctioning malevolent machines so beloved of the makers of blockbuster movies. Crazed intelligent machines would not, Hollywood has warned up, feel bound by the Geneva convention, nor indeed by any notion of rights for mere humans.

The Russians are not alone in this. Experts reporting to the UN rights body say the US, UK, South Korea, and Japan are all working on "lethal autonomous robotics". And you thought drones were bad …