The DeepMind team of Google has been trying to create a way while working along with scientists so that before the artificial intelligence can turn on us, they can be killed. Considering the fact that the tech giants are developing AI systems which are said to be more capable and intelligent, this is something that can be a matter of concern, however it seems to be that Google has it covered.

In 2014, DeepMind was acquired by Google for around $400 million. Ever since then, a lot of incredible things have been done by the company. The recent achievement of the company was the five game matches where Lee Sedol, the highest ranked Go Player was beaten with a score of 4‐1. DeepMind not only believes in great technologies but also believe in becoming greater in what they do and always believe in learning. The creators are now worried as to what could be instore if it is decided by AI‘s that they do not want to be controlled by humans anymore.

Safety Interruptible Agents it the title that is on Paper which has plans that have been revealed by Oxford University and DeepMind scientists on this front. If the real time operation of such an agent is under human supervision, then it is very important that the big red button should be pressed by the human operator. This will help to present the agent from doing any kind of action that can be harmful. This can be harmful not just for the environment but also for the agent. It has also been explained on the paper, that with the big red button, it will help ensure that the agent is lead into a situation that is safer. It has also been confirmed on paper that a framework is in place which will allow the human to intervene safely with the AI as well as ensure that the interruptions are prevented.