Artificial Intelligence has been an emerging field over the past couple of decades and while we’re not quite close to having our own home-running JARVIS AIs, the field has come a long way. Due to advancements in the area we were given the likes of Siri and Cortana, and is even instrumental to changes in the SEO industry. But what happens when one day, Arnold Schwarzenegger has had enough and goes berserk due to an AI. Due to the slight possibility of that happening, Google’s Deep Mind team is now building a kill switch for AIs.
The team recently published a paper outlining how exactly it means to do this. This particular paper, titled “Safely Interruptible Agents,” investigates how to turn off AI if it starts doing something its human operator doesn’t want it to do, such as sending robots to the past to kill certain human beings. Technical jargon aside, the Deep Mind team have developed a framework that’ll keep AI from learning how to prevent — or induce — human interruption of whatever it’s doing.
“We have proposed a framework to allow a human operator to repeatedly safely interrupt a reinforcement learning agent while making sure the agent will not learn to prevent or induce these interruptions.”
They theorised a situation where a robot was working in a warehouse, sorting boxes or going outside to bring more boxes in. The latter is considered more important, so the researchers would give the robot a bigger reward for doing so. But human intervention to prevent damage is needed because it rains pretty frequently here. That alters the task for the robot, making it want to stay out of the rain, and then adopting the human interruption as part of the task rather than being a one-off thing.
Source: Business Insider
