'Robot uprising? It's miles off', says cybersecurity advisor of Google's kill-switch plans

Laurent Orseau from Google DeepMind, and Stuart Armstrong from the Future of Humanity Institute at the university outlined how future intelligent machines could be prevented from learning to override human commands with specific coding.

"So, I'll just empty the tray and then take over the world"

Thursday, June 9, 2016

From The Terminator to The Matrix, Hollywood has long held a fascination with the idea of robots rising up against humanity. But while tech giants Google have this week announced development of a 'kill switch' to prevent that nightmare scenario taking place, cybersecurity advisor Graeme Batsman has told talkRADIO that mechanical world domination remains "miles off". 

Scientists from Google's research division, known as DeepMind, have paired up with Oxford University's brilliant named Future of Humanity Institute to develop technology that will shut down any kind of rebellion from any form of artificial intelligence (AI).

The research is centred on a method to ensure AIs, which learn via reinforcement, cannot develop mechanisms which allow them to avoid being over-ruled by human managers.

Despite the serious nature of the research, Batsman, a senior cybersecurity specialist, reassured listeners that the likelihood of a robot uprising remains firmly in science fiction. 

"Computers are kind of stupid, you have to tell them to do something," he told Julia Hartley-Brewer. "They can't read your mind and write their own code base

"It's good they're thinking of [the kill switch] now, because security's often an afterthought, but I think any dangers of this are many miles off."